Re: Status of the t64 transition

[Andreas Metzler]

On 2024-04-18 Sebastian Ramacher  wrote:
[...]
> Let's start with the first category. Those are packages that could be
> binNMUed, but there are issues that make those rebuilds not have the
> desired effect. This list include packages that
>  * are BD-Uninstallabe,
>  * FTBFS but with out ftbfs-tagged RC bug,
>  * have hard-coded dependencies on pre-t64 libraries,
>  * have $oldlib | $newlib dependencies (those are at least wrong on
>armel/armhf and violate policy 2.2.1 once the pre-t64 libraries are
>decrufted),
>  * have been rebuilt before all dependencies were built,
>  * have broken symbols/shlibs files producing incorrect dependencies,
>  * or might just be missing the binNMU (but those should be few).

> hugin
[...]

Good morning,

thanks for the update.

Looking at hugin, I think it is fine on all release-architectures, none
of the problems noted above apply here. Am I missing something?

TIA, cu Andreas

PS: fakeroot seems to be an important blocker not in the list.

Bug#1067729: nmu: exim4_4.97-5

[Andreas Metzler]

On 2024-03-26 Andreas Metzler  wrote:
[...]
> nmu exim4_4.97-5 . armel armhf hppa m68k . unstable . -m "Rebuild against 
> libspf2-dev >= 1.2.10-8.1 (64-bit time_t transition)"

> The first t64-changed libspf2 was uninstallable on the 32bit archs,
> which is why exim4 was not bin-nmued successfully there yet. This is
> fixed now.

> This can only be done successfully after libtirpc 1.3.4+ds-1.2 has
> passed NEW processing.

libtirpc has been accepted. :-)

The exim4 changelog entry should refer to -8.2, though:

nmu exim4_4.97-5 . armel armhf hppa m68k . unstable . -m "Rebuild against 
libspf2-dev 1.2.10-8.2 (64-bit time_t transition)"

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#1067729: nmu: exim4_4.97-5

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Control: affects -1 + src:exim4
User: release.debian@packages.debian.org
Usertags: binnmu

Hello,

nmu exim4_4.97-5 . armel armhf hppa m68k . unstable . -m "Rebuild against 
libspf2-dev >= 1.2.10-8.1 (64-bit time_t transition)"

The first t64-changed libspf2 was uninstallable on the 32bit archs,
which is why exim4 was not bin-nmued successfully there yet. This is
fixed now.

This can only be done successfully after libtirpc 1.3.4+ds-1.2 has
passed NEW processing.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#1061190: bullseye-pu: package gnutls28/3.7.1-5+deb11u5

[Andreas Metzler]

On 2024-03-01 Salvatore Bonaccorso  wrote:
> On Thu, Feb 01, 2024 at 06:35:38AM +, Adam D. Barratt wrote:
>> Control: tags -1 + confirmed
 
>> On Sat, 2024-01-20 at 15:53 +0100, Andreas Metzler wrote:
>>> I would like to fix both CVE-2024-0567 and CVE-2024-0553 via a
>>> oldstable-updates since they do not require a DSA.

>> Please go ahead.

> Andreas did you saw the ack from Adam?

> FTR, please keep the CVE references now as we have the incomplete fix
> in bullseye for CVE-2023-5981 with the 3.7.1-5+deb11u4 .

Good Morning,

Thank you for the reminder, looks like I only did a "dput -s" instead of
the actual dput.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#1061190: bullseye-pu: package gnutls28/3.7.1-5+deb11u5

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bullseye
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: gnutl...@packages.debian.org, t...@security.debian.org
Control: affects -1 + src:gnutls28

Hello,

I would like to fix both CVE-2024-0567 and CVE-2024-0553 via a
oldstable-updates since they do not require a DSA.

[ Checklist ]
  [x] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
  [x] the issue is verified as fixed in unstable

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.7.1/debian/changelog gnutls28-3.7.1/debian/changelog
--- gnutls28-3.7.1/debian/changelog	2023-11-30 11:37:44.0 +0100
+++ gnutls28-3.7.1/debian/changelog	2024-01-20 07:56:15.0 +0100
@@ -1,3 +1,13 @@
+gnutls28 (3.7.1-5+deb11u5) bullseye; urgency=medium
+
+  * Cherrypick two CVE fixes from 3.8.3:
+Fix assertion failure when verifying a certificate chain with a cycle of
+cross signatures. CVE-2024-0567 GNUTLS-SA-2024-01-09 Closes: #1061045
+Fix more timing side-channel inside RSA-PSK key exchange. CVE-2024-0553
+GNUTLS-SA-2024-01-14 Closes: #1061046
+
+ -- Andreas Metzler   Sat, 20 Jan 2024 07:56:15 +0100
+
 gnutls28 (3.7.1-5+deb11u4) bullseye; urgency=medium
 
   * Backport fix for CVE-2023-5981 / GNUTLS-SA-2023-10-23 (timing sidechannel
diff -Nru gnutls28-3.7.1/debian/patches/63-x509-detect-loop-in-certificate-chain.patch gnutls28-3.7.1/debian/patches/63-x509-detect-loop-in-certificate-chain.patch
--- gnutls28-3.7.1/debian/patches/63-x509-detect-loop-in-certificate-chain.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.7.1/debian/patches/63-x509-detect-loop-in-certificate-chain.patch	2024-01-20 07:56:15.0 +0100
@@ -0,0 +1,188 @@
+From 9edbdaa84e38b1bfb53a7d72c1de44f8de373405 Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Thu, 11 Jan 2024 15:45:11 +0900
+Subject: [PATCH 1/2] x509: detect loop in certificate chain
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+There can be a loop in a certificate chain, when multiple CA
+certificates are cross-signed with each other, such as A → B, B → C,
+and C → A.  Previously, the verification logic was not capable of
+handling this scenario while sorting the certificates in the chain in
+_gnutls_sort_clist, resulting in an assertion failure.  This patch
+properly detects such loop and aborts further processing in a graceful
+manner.
+
+Signed-off-by: Daiki Ueno 
+---
+ lib/x509/common.c   |   4 ++
+ tests/test-chains.h | 125 
+ 2 files changed, 129 insertions(+)
+
+--- a/lib/x509/common.c
 b/lib/x509/common.c
+@@ -1794,10 +1794,14 @@ unsigned int _gnutls_sort_clist(gnutls_x
+ 		prev = issuer[prev];
+ 		if (prev < 0) {	/* no issuer */
+ 			break;
+ 		}
+ 
++		if (insorted[prev]) { /* loop detected */
++			break;
++		}
++
+ 		sorted[i] = clist[prev];
+ 		insorted[prev] = 1;
+ 	}
+ 
+ 	/* append the remaining certs */
+--- a/tests/test-chains.h
 b/tests/test-chains.h
+@@ -4261,10 +4261,133 @@ static const char *rsa_sha1_not_in_trust
+ 	"tnYFXKC0Q+QUf38horqG2Mc3/uh8MOm0eYUXwGJOdXYD\n"
+ 	"-END CERTIFICATE-\n",
+ 	NULL
+ };
+ 
++static const char *cross_signed[] = {
++	/* server (signed by A1) */
++	"-BEGIN CERTIFICATE-\n"
++	"MIIBqDCCAVqgAwIBAgIUejlil+8DBffazcnMNwyOOP6yCCowBQYDK2VwMBoxGDAW\n"
++	"BgNVBAMTD0ludGVybWVkaWF0ZSBBMTAgFw0yNDAxMTEwNjI3MjJaGA85OTk5MTIz\n"
++	"MTIzNTk1OVowNzEbMBkGA1UEChMSR251VExTIHRlc3Qgc2VydmVyMRgwFgYDVQQD\n"
++	"Ew90ZXN0LmdudXRscy5vcmcwKjAFBgMrZXADIQA1ZVS0PcNeTPQMZ+FuVz82AHrj\n"
++	"qL5hWEpCDgpG4M4fxaOBkjCBjzAMBgNVHRMBAf8EAjAAMBoGA1UdEQQTMBGCD3Rl\n"
++	"c3QuZ251dGxzLm9yZzATBgNVHSUEDDAKBggrBgEFBQcDATAOBgNVHQ8BAf8EBAMC\n"
++	"B4AwHQYDVR0OBBYEFGtEUv+JSt+zPoO3lu0IiObZVoiNMB8GA1UdIwQYMBaAFPnY\n"
++	"v6Pw0IvKSqIlb6ewHyEAmTA3MAUGAytlcANBAAS2lyc87kH/aOvNKzPjqDwUYxPA\n"
++	"CfYjyaKea2d0DZLBM5+Bjnj/4aWwTKgVTJzWhLJcLtaSdVHrXqjr9NhEhQ0=\n"
++	"-END CERTIFICATE-\n",
++	/* A1 (signed by A) */
++	"-BEGIN CERTIFICATE-\n"
++	"MIIBUjCCAQSgAwIBAgIUe/R+NVp04e74ySw2qgI6KZgFR20wBQYDK2VwMBExDzAN\n"
++	"BgNVBAMTBlJvb3QgQTAgFw0yNDAxMTEwNjI1MDFaGA85OTk5MTIzMTIzNTk1OVow\n"
++	"GjEYMBYGA1UEAxMPSW50ZXJtZWRpYXRlIEExMCowBQYDK2VwAyEAlkTNqwz973sy\n"
++	"u3whMjSiUMs77CZu5YA7Gi5KcakExrKjYzBhMA8GA1UdEwEB/wQFMAMBAf8wDgYD\n"
++	"VR0PAQH/BAQDAgIEMB0GA1UdDgQWBBT52L+j8NCLykqiJW+nsB8hAJkwNzAfBgNV\n"
++	"HSMEGDAWgBRbYgOkRGsd3Z74+CauX4htzLg0lzAFBgMrZXADQQBM0NBaFVPd3cTJ\n"
++	"DSaZNT34fsHuJk4eagpn8mBxKQpghq4s8Ap+nYtp

Bug#1061189: bookworm-pu: package gnutls28/3.7.9-2+deb12u2

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: gnutl...@packages.debian.org, t...@security.debian.org, 
gnutl...@packages.debian.org
Control: affects -1 + src:gnutls28

Hello,

I would like to fix both CVE-2024-0567 and CVE-2024-0553 via a
stable-updates since they do not require a DSA.

[ Checklist ]
  [x] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
  [x] the issue is verified as fixed in unstable

cu Andreas
diff -Nru gnutls28-3.7.9/debian/changelog gnutls28-3.7.9/debian/changelog
--- gnutls28-3.7.9/debian/changelog	2023-11-30 07:50:48.0 +0100
+++ gnutls28-3.7.9/debian/changelog	2024-01-19 18:28:37.0 +0100
@@ -1,3 +1,13 @@
+gnutls28 (3.7.9-2+deb12u2) bookworm; urgency=medium
+
+  * Cherrypick two CVE fixes from 3.8.3:
+Fix assertion failure when verifying a certificate chain with a cycle of
+cross signatures. CVE-2024-0567 GNUTLS-SA-2024-01-09 Closes: #1061045
+Fix more timing side-channel inside RSA-PSK key exchange. CVE-2024-0553
+GNUTLS-SA-2024-01-14 Closes: #1061046
+
+ -- Andreas Metzler   Fri, 19 Jan 2024 18:28:37 +0100
+
 gnutls28 (3.7.9-2+deb12u1) bookworm; urgency=medium
 
   * Backport fix for CVE-2023-5981 / GNUTLS-SA-2023-10-23 (timing sidechannel
diff -Nru gnutls28-3.7.9/debian/patches/61-x509-detect-loop-in-certificate-chain.patch gnutls28-3.7.9/debian/patches/61-x509-detect-loop-in-certificate-chain.patch
--- gnutls28-3.7.9/debian/patches/61-x509-detect-loop-in-certificate-chain.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.7.9/debian/patches/61-x509-detect-loop-in-certificate-chain.patch	2024-01-19 18:28:07.0 +0100
@@ -0,0 +1,188 @@
+From 9edbdaa84e38b1bfb53a7d72c1de44f8de373405 Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Thu, 11 Jan 2024 15:45:11 +0900
+Subject: [PATCH 1/2] x509: detect loop in certificate chain
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+There can be a loop in a certificate chain, when multiple CA
+certificates are cross-signed with each other, such as A → B, B → C,
+and C → A.  Previously, the verification logic was not capable of
+handling this scenario while sorting the certificates in the chain in
+_gnutls_sort_clist, resulting in an assertion failure.  This patch
+properly detects such loop and aborts further processing in a graceful
+manner.
+
+Signed-off-by: Daiki Ueno 
+---
+ lib/x509/common.c   |   4 ++
+ tests/test-chains.h | 125 
+ 2 files changed, 129 insertions(+)
+
+--- a/lib/x509/common.c
 b/lib/x509/common.c
+@@ -1794,10 +1794,14 @@ unsigned int _gnutls_sort_clist(gnutls_x
+ 		prev = issuer[prev];
+ 		if (prev < 0) {	/* no issuer */
+ 			break;
+ 		}
+ 
++		if (insorted[prev]) { /* loop detected */
++			break;
++		}
++
+ 		sorted[i] = clist[prev];
+ 		insorted[prev] = 1;
+ 	}
+ 
+ 	/* append the remaining certs */
+--- a/tests/test-chains.h
 b/tests/test-chains.h
+@@ -4261,10 +4261,133 @@ static const char *rsa_sha1_not_in_trust
+ 	"tnYFXKC0Q+QUf38horqG2Mc3/uh8MOm0eYUXwGJOdXYD\n"
+ 	"-END CERTIFICATE-\n",
+ 	NULL
+ };
+ 
++static const char *cross_signed[] = {
++	/* server (signed by A1) */
++	"-BEGIN CERTIFICATE-\n"
++	"MIIBqDCCAVqgAwIBAgIUejlil+8DBffazcnMNwyOOP6yCCowBQYDK2VwMBoxGDAW\n"
++	"BgNVBAMTD0ludGVybWVkaWF0ZSBBMTAgFw0yNDAxMTEwNjI3MjJaGA85OTk5MTIz\n"
++	"MTIzNTk1OVowNzEbMBkGA1UEChMSR251VExTIHRlc3Qgc2VydmVyMRgwFgYDVQQD\n"
++	"Ew90ZXN0LmdudXRscy5vcmcwKjAFBgMrZXADIQA1ZVS0PcNeTPQMZ+FuVz82AHrj\n"
++	"qL5hWEpCDgpG4M4fxaOBkjCBjzAMBgNVHRMBAf8EAjAAMBoGA1UdEQQTMBGCD3Rl\n"
++	"c3QuZ251dGxzLm9yZzATBgNVHSUEDDAKBggrBgEFBQcDATAOBgNVHQ8BAf8EBAMC\n"
++	"B4AwHQYDVR0OBBYEFGtEUv+JSt+zPoO3lu0IiObZVoiNMB8GA1UdIwQYMBaAFPnY\n"
++	"v6Pw0IvKSqIlb6ewHyEAmTA3MAUGAytlcANBAAS2lyc87kH/aOvNKzPjqDwUYxPA\n"
++	"CfYjyaKea2d0DZLBM5+Bjnj/4aWwTKgVTJzWhLJcLtaSdVHrXqjr9NhEhQ0=\n"
++	"-END CERTIFICATE-\n",
++	/* A1 (signed by A) */
++	"-BEGIN CERTIFICATE-\n"
++	"MIIBUjCCAQSgAwIBAgIUe/R+NVp04e74ySw2qgI6KZgFR20wBQYDK2VwMBExDzAN\n"
++	"BgNVBAMTBlJvb3QgQTAgFw0yNDAxMTEwNjI1MDFaGA85OTk5MTIzMTIzNTk1OVow\n"
++	"GjEYMBYGA1UEAxMPSW50ZXJtZWRpYXRlIEExMCowBQYDK2VwAyEAlkTNqwz973sy\n"
++	"u3whMjSiUMs77CZu5YA7Gi5KcakExrKjYzBhMA8GA1UdEwEB/wQFMAMBAf8wDgYD\n"
++	"VR0PAQH/BAQDAgIEMB0GA1UdDgQWBBT52L+j8NCLykqiJW+nsB8hAJkwNzAfBgNV\n"
++	"HSMEGDAWgBRbYgOkRGsd3Z74+CauX4htzLg0lzAFBgMrZXADQQBM0NBaFVPd3cTJ\n"
++	"DSaZNT34fsHuJk4eagpn8mBxKQpghq4s8Ap+nYtp2KiXjcizss53PeLXVnkfyLi0\n"
++	"TLVBHvUJ\n"
++	"-END CERTIFICATE-\n",
++	/* A (signed by B) */
++	"-BE

Bug#1057137: bullseye-pu: package gnutls28/3.7.1-5+deb11u4

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bullseye
User: release.debian@packages.debian.org
Usertags: pu
Control: affects -1 + src:gnutls28

Hello,

I would like to fix CVE-2023-5981 / GNUTLS-SA-2023-10-23 for oldstable
(no DSA forthcoming, to fixed by stable update.) The patch is
cherrypicked from upstream 3.8.2 release.

[ Checklist ]
  [x] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
  [x] the issue is verified as fixed in unstable

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.7.1/debian/changelog gnutls28-3.7.1/debian/changelog
--- gnutls28-3.7.1/debian/changelog	2023-02-12 13:59:45.0 +0100
+++ gnutls28-3.7.1/debian/changelog	2023-11-30 11:37:44.0 +0100
@@ -1,3 +1,10 @@
+gnutls28 (3.7.1-5+deb11u4) bullseye; urgency=medium
+
+  * Backport fix for CVE-2023-5981 / GNUTLS-SA-2023-10-23 (timing sidechannel
+in RSA-PSK key exchange) from 3.8.2. Closes: #1056188
+
+ -- Andreas Metzler   Thu, 30 Nov 2023 11:37:44 +0100
+
 gnutls28 (3.7.1-5+deb11u3) bullseye-security; urgency=high
 
   * Fix timing sidechannel vulnerability in RSA decryption.
diff -Nru gnutls28-3.7.1/debian/patches/62-auth-rsa_psk-side-step-potential-side-channel.patch gnutls28-3.7.1/debian/patches/62-auth-rsa_psk-side-step-potential-side-channel.patch
--- gnutls28-3.7.1/debian/patches/62-auth-rsa_psk-side-step-potential-side-channel.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.7.1/debian/patches/62-auth-rsa_psk-side-step-potential-side-channel.patch	2023-11-30 11:37:44.0 +0100
@@ -0,0 +1,229 @@
+From 29d6298d0b04cfff970b993915db71ba3f580b6d Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Mon, 23 Oct 2023 09:26:57 +0900
+Subject: [PATCH] auth/rsa_psk: side-step potential side-channel
+
+This removes branching that depends on secret data, porting changes
+for regular RSA key exchange from
+4804febddc2ed958e5ae774de2a8f85edeeff538 and
+80a6ce8ddb02477cd724cd5b2944791aaddb702a.  This also removes the
+allow_wrong_pms as it was used sorely to control debug output
+depending on the branching.
+
+Signed-off-by: Daiki Ueno 
+---
+ lib/auth/rsa.c |  2 +-
+ lib/auth/rsa_psk.c | 90 ++
+ lib/gnutls_int.h   |  4 ---
+ lib/priority.c |  1 -
+ 4 files changed, 35 insertions(+), 62 deletions(-)
+
+--- a/lib/auth/rsa.c
 b/lib/auth/rsa.c
+@@ -205,11 +205,11 @@ proc_rsa_client_kx(gnutls_session_t sess
+ 	gnutls_privkey_decrypt_data2(session->internals.selected_key,
+  0, , session->key.key.data,
+  session->key.key.size);
+ 	/* After this point, any conditional on failure that cause differences
+ 	 * in execution may create a timing or cache access pattern side
+-	 * channel that can be used as an oracle, so treat very carefully */
++	 * channel that can be used as an oracle, so tread carefully */
+ 
+ 	/* Error handling logic:
+ 	 * In case decryption fails then don't inform the peer. Just use the
+ 	 * random key previously generated. (in order to avoid attack against
+ 	 * pkcs-1 formatting).
+--- a/lib/auth/rsa_psk.c
 b/lib/auth/rsa_psk.c
+@@ -262,18 +262,17 @@ static int
+ _gnutls_proc_rsa_psk_client_kx(gnutls_session_t session, uint8_t * data,
+ 			   size_t _data_size)
+ {
+ 	gnutls_datum_t username;
+ 	psk_auth_info_t info;
+-	gnutls_datum_t plaintext;
+ 	gnutls_datum_t ciphertext;
+ 	gnutls_datum_t pwd_psk = { NULL, 0 };
+ 	int ret, dsize;
+-	int randomize_key = 0;
+ 	ssize_t data_size = _data_size;
+ 	gnutls_psk_server_credentials_t cred;
+ 	gnutls_datum_t premaster_secret = { NULL, 0 };
++	volatile uint8_t ver_maj, ver_min;
+ 
+ 	cred = (gnutls_psk_server_credentials_t)
+ 	_gnutls_get_cred(session, GNUTLS_CRD_PSK);
+ 
+ 	if (cred == NULL) {
+@@ -325,75 +324,53 @@ _gnutls_proc_rsa_psk_client_kx(gnutls_se
+ 		gnutls_assert();
+ 		return GNUTLS_E_UNEXPECTED_PACKET_LENGTH;
+ 	}
+ 	ciphertext.size = dsize;
+ 
+-	ret =
+-	gnutls_privkey_decrypt_data(session->internals.selected_key, 0,
+-	, );
+-	if (ret < 0 || plaintext.size != GNUTLS_MASTER_SIZE) {
+-		/* In case decryption fails then don't inform
+-		 * the peer. Just use a random key. (in order to avoid
+-		 * attack against pkcs-1 formatting).
+-		 */
+-		gnutls_assert();
+-		_gnutls_debug_log
+-		("auth_rsa_psk: Possible PKCS #1 format attack\n");
+-		if (ret >= 0) {
+-			gnutls_free(plaintext.data);
+-		}
+-		randomize_key = 1;
+-	} else {
+-		/* If the secret was properly formatted, then
+-		 * check the version number.
+-		 */
+-		if (_gnutls_get_adv_version_major(session) !=
+-		plaintext.data[0]
+-		|| (session->internals.allow_wrong_pms == 0
+-			&& _gnutls_get_adv_version_minor(session) !=
+-			plaintext.data[1])) {
+-			/* No error is returned here, if the version

Bug#1057128: bookworm-pu: package gnutls28/3.7.9-2+deb12u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
Control: affects -1 + src:gnutls28

Hello,

I would like to fix CVE-2023-5981 / GNUTLS-SA-2023-10-23 for stable (no
DSA forthcoming, to fixed by stable update.) The patch is cherrypicked
from upstream 3.8.2 release. Ubuntu's 3.7.8-5ubuntu1.1 has the same
patch (except for being U3 instead of U5 format).

[ Checklist ]
  [x] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
  [x] the issue is verified as fixed in unstable

cu andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.7.9/debian/changelog gnutls28-3.7.9/debian/changelog
--- gnutls28-3.7.9/debian/changelog	2023-04-15 13:45:57.0 +0200
+++ gnutls28-3.7.9/debian/changelog	2023-11-30 07:50:48.0 +0100
@@ -1,3 +1,10 @@
+gnutls28 (3.7.9-2+deb12u1) bookworm; urgency=medium
+
+  * Backport fix for CVE-2023-5981 / GNUTLS-SA-2023-10-23 (timing sidechannel
+in RSA-PSK key exchange) from 3.8.2. Closes: #1056188
+
+ -- Andreas Metzler   Thu, 30 Nov 2023 07:50:48 +0100
+
 gnutls28 (3.7.9-2) unstable; urgency=medium
 
   * CI: Do not try to run tests/ktls.sh, it uses a helper binary. (Plus gnutls
diff -Nru gnutls28-3.7.9/debian/patches/60-auth-rsa_psk-side-step-potential-side-channel.patch gnutls28-3.7.9/debian/patches/60-auth-rsa_psk-side-step-potential-side-channel.patch
--- gnutls28-3.7.9/debian/patches/60-auth-rsa_psk-side-step-potential-side-channel.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.7.9/debian/patches/60-auth-rsa_psk-side-step-potential-side-channel.patch	2023-11-30 07:50:48.0 +0100
@@ -0,0 +1,229 @@
+From 29d6298d0b04cfff970b993915db71ba3f580b6d Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Mon, 23 Oct 2023 09:26:57 +0900
+Subject: [PATCH] auth/rsa_psk: side-step potential side-channel
+
+This removes branching that depends on secret data, porting changes
+for regular RSA key exchange from
+4804febddc2ed958e5ae774de2a8f85edeeff538 and
+80a6ce8ddb02477cd724cd5b2944791aaddb702a.  This also removes the
+allow_wrong_pms as it was used sorely to control debug output
+depending on the branching.
+
+Signed-off-by: Daiki Ueno 
+---
+ lib/auth/rsa.c |  2 +-
+ lib/auth/rsa_psk.c | 90 ++
+ lib/gnutls_int.h   |  4 ---
+ lib/priority.c |  1 -
+ 4 files changed, 35 insertions(+), 62 deletions(-)
+
+--- a/lib/auth/rsa.c
 b/lib/auth/rsa.c
+@@ -205,11 +205,11 @@ proc_rsa_client_kx(gnutls_session_t sess
+ 	gnutls_privkey_decrypt_data2(session->internals.selected_key,
+  0, , session->key.key.data,
+  session->key.key.size);
+ 	/* After this point, any conditional on failure that cause differences
+ 	 * in execution may create a timing or cache access pattern side
+-	 * channel that can be used as an oracle, so treat very carefully */
++	 * channel that can be used as an oracle, so tread carefully */
+ 
+ 	/* Error handling logic:
+ 	 * In case decryption fails then don't inform the peer. Just use the
+ 	 * random key previously generated. (in order to avoid attack against
+ 	 * pkcs-1 formatting).
+--- a/lib/auth/rsa_psk.c
 b/lib/auth/rsa_psk.c
+@@ -262,18 +262,17 @@ static int
+ _gnutls_proc_rsa_psk_client_kx(gnutls_session_t session, uint8_t * data,
+ 			   size_t _data_size)
+ {
+ 	gnutls_datum_t username;
+ 	psk_auth_info_t info;
+-	gnutls_datum_t plaintext;
+ 	gnutls_datum_t ciphertext;
+ 	gnutls_datum_t pwd_psk = { NULL, 0 };
+ 	int ret, dsize;
+-	int randomize_key = 0;
+ 	ssize_t data_size = _data_size;
+ 	gnutls_psk_server_credentials_t cred;
+ 	gnutls_datum_t premaster_secret = { NULL, 0 };
++	volatile uint8_t ver_maj, ver_min;
+ 
+ 	cred = (gnutls_psk_server_credentials_t)
+ 	_gnutls_get_cred(session, GNUTLS_CRD_PSK);
+ 
+ 	if (cred == NULL) {
+@@ -327,75 +326,53 @@ _gnutls_proc_rsa_psk_client_kx(gnutls_se
+ 		gnutls_assert();
+ 		return GNUTLS_E_UNEXPECTED_PACKET_LENGTH;
+ 	}
+ 	ciphertext.size = dsize;
+ 
+-	ret =
+-	gnutls_privkey_decrypt_data(session->internals.selected_key, 0,
+-	, );
+-	if (ret < 0 || plaintext.size != GNUTLS_MASTER_SIZE) {
+-		/* In case decryption fails then don't inform
+-		 * the peer. Just use a random key. (in order to avoid
+-		 * attack against pkcs-1 formatting).
+-		 */
+-		gnutls_assert();
+-		_gnutls_debug_log
+-		("auth_rsa_psk: Possible PKCS #1 format attack\n");
+-		if (ret >= 0) {
+-			gnutls_free(plaintext.data);
+-		}
+-		randomize_key = 1;
+-	} else {
+-		/* If the secret was properly formatted, then
+-		 * check the version number.
+-		 */
+-		if (_gnutls_get_adv_version_major(session) !=
+-		plaintext.data[0]
+-		|| (session->internals.allow_wrong_pms == 0
+-			&& _gnutls_get_adv_version_

Bug#1055155: bookworm-pu: package exim4/4.96-15+deb12u3 (2nd try for new bug)

[Andreas Metzler]

On 2023-11-04 Andreas Metzler  wrote:
[...]
> Thank you, updated.

Another iteration, adding 
 + 76-14-Lookups-Fix-dnsdb-lookup-of-multi-chunk-TXT.-Bug-305.patch Fix
   regression in dnsdb in CVE-2023-42119 fix. (Upstream bug 3054)

dnsdb lookups were swallowing the last character of multiline TXT
records, as shown with 
/usr/sbin/exim4 -be '${lookup dnsdb{>\n; txt=exim.org}}'
outputting "cumi" instead of "cumin"

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-09-29 22:38:02.0 +0200
+++ exim4-4.96/debian/changelog	2023-11-18 11:07:57.0 +0100
@@ -1,3 +1,31 @@
+exim4 (4.96-15+deb12u3) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
++ 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
++ 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
++ 76-12-DNS-more-hardening-against-crafted-responses.patch
++ 76-14-Lookups-Fix-dnsdb-lookup-of-multi-chunk-TXT.-Bug-305.patch Fix
+  regression in dnsdb in CVE-2023-42119 fix. (Upstream bug 3054)
+  * tests/basic: Add isolation-container restriction (needs a running
+exim daemon).
+  * Add ${run } expansion test to tests/basic.
+  * Update code to 4.96.2, fixing issues with the proxy protocol
+(CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42119). It
+also includes additional hardening for spf lookups, however CVE-2023-42118
+was diagnosed as a vulnerability in the libspf2 library and needs to be
+addressed there. Closes: #1053310
+
+ -- Andreas Metzler   Wed, 18 Nov 2023 11:07:57 +0100
+
 exim4 (4.96-15+deb12u2) bookworm-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-11-18 11:07:57.0 +0100
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actual EHLO LIMITS values with the
+ cached values and invalidate cache if different.  OK to carry on with
+ connect since values are advisory. */
+-- 
+2.40.1
+
diff -Nru exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
--- exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	2023-11-18 11:07:57.0 +0100
@@ -0,0 +1,99 @@
+From 1209e3e19e292cee517e43a2ccfe9b44b33bb1dc Mon Sep 17 00:00:00 2001
+From: Jasen Betts 
+Date: Sun, 23 Jul 2023 13:43:59 +0100
+Subject: [PATCH] Expansions: disallow UTF-16 surrogates from ${utf8clean:...}.
+  Bug 2998
+
+---
+ doc/ChangeLog |  4 
+ src/expand.c  | 27 +--
+ 2 files changed, 21 insertions(+), 10 deletions(-)
+
+--- a/src/expand.c
 b/src/expand.c
+@@ -7731,11 +7731,11 @@ NOT_ITEM: ;
+ 
+ 	case EOP_UTF8CLEAN:
+ 	  {
+ 	  int seq_len = 0, index = 0;
+ 	  int by

Bug#1055155: bookworm-pu: package exim4/4.96-15+deb12u3 (2nd try for new bug)

[Andreas Metzler]

On 2023-11-04 Salvatore Bonaccorso  wrote:
> On Wed, Nov 01, 2023 at 12:03:37PM +0100, Andreas Metzler wrote:
[...]
> > * Update code to 4.96.2, fixing issues with the proxy protocol
> >   (CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42219). It
> >   also includes additional hardening for spf lookups, however CVE-2023-42218

> The mentioned CVEs have a typo. I believe this should be
> CVE-2023-42117 and CVE-2023-42119 (and for completeness about the
> libspf2 mentioning CVE-2023-42118).

Thank you, updated.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-09-29 22:38:02.0 +0200
+++ exim4-4.96/debian/changelog	2023-11-01 07:07:57.0 +0100
@@ -1,3 +1,29 @@
+exim4 (4.96-15+deb12u3) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
++ 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
++ 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
++ 76-12-DNS-more-hardening-against-crafted-responses.patch
+  * tests/basic: Add isolation-container restriction (needs a running
+exim daemon).
+  * Add ${run } expansion test to tests/basic.
+  * Update code to 4.96.2, fixing issues with the proxy protocol
+(CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42119). It
+also includes additional hardening for spf lookups, however CVE-2023-42118
+was diagnosed as a vulnerability in the libspf2 library and needs to be
+addressed there. Closes: #1053310
+
+ -- Andreas Metzler   Wed, 01 Nov 2023 07:07:57 +0100
+
 exim4 (4.96-15+deb12u2) bookworm-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-11-01 07:03:21.0 +0100
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actual EHLO LIMITS values with the
+ cached values and invalidate cache if different.  OK to carry on with
+ connect since values are advisory. */
+-- 
+2.40.1
+
diff -Nru exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
--- exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	2023-11-01 07:03:21.0 +0100
@@ -0,0 +1,99 @@
+From 1209e3e19e292cee517e43a2ccfe9b44b33bb1dc Mon Sep 17 00:00:00 2001
+From: Jasen Betts 
+Date: Sun, 23 Jul 2023 13:43:59 +0100
+Subject: [PATCH] Expansions: disallow UTF-16 surrogates from ${utf8clean:...}.
+  Bug 2998
+
+---
+ doc/ChangeLog |  4 
+ src/expand.c  | 27 +--
+ 2 files changed, 21 insertions(+), 10 deletions(-)
+
+--- a/src/expand.c
 b/src/expand.c
+@@ -7731,11 +7731,11 @@ NOT_ITEM: ;
+ 
+ 	case EOP_UTF8CLEAN:
+ 	  {
+ 	  int seq_len = 0, index = 0;
+ 	  int bytes_left = 0;
+-	  long codepoint = -1;

Bug#1055155: bookworm-pu: package exim4/4.96-15+deb12u3 (2nd try for new bug)

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
Control: affects -1 + src:exim4

Hello,

I would like to push another round of cherry-picked upstream fixes to
bookworm, including the update to 4.96.2 to fix two non-DSA minor
security issues.

The changes are included in the new upstream (4.97 rc) uploads to sid which=
 are present in sid and testing.


* Multiple bugfixes from upstream GIT master:
  + 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
  + 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
(Upstream bug 2998)
  + 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
  + 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
(Upstream bug 3013)
> ${run expansion breakage, similar to #1025420.
  + 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
TLS cert expiry date. Closes: #1043233
(Upstream bug 3014)
> This is major hickup, bordering on RC.

  + 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
> Another patch for ${run} expansion breakage.
  + 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
  + 76-12-DNS-more-hardening-against-crafted-responses.patch
* tests/basic: Add isolation-container restriction (needs a running
  exim daemon).
* Add ${run } expansion test to tests/basic.
* Update code to 4.96.2, fixing issues with the proxy protocol
  (CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42219). It
  also includes additional hardening for spf lookups, however CVE-2023-42218
  was diagnosed as a vulnerability in the libspf2 library and needs to be
  addressed there. Closes: #1053310

cu Andreas
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-09-29 22:38:02.0 +0200
+++ exim4-4.96/debian/changelog	2023-11-01 07:07:57.0 +0100
@@ -1,3 +1,29 @@
+exim4 (4.96-15+deb12u3) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
++ 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
++ 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
++ 76-12-DNS-more-hardening-against-crafted-responses.patch
+  * tests/basic: Add isolation-container restriction (needs a running
+exim daemon).
+  * Add ${run } expansion test to tests/basic.
+  * Update code to 4.96.2, fixing issues with the proxy protocol
+(CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42219). It
+also includes additional hardening for spf lookups, however CVE-2023-42218
+was diagnosed as a vulnerability in the libspf2 library and needs to be
+addressed there. Closes: #1053310
+
+ -- Andreas Metzler   Wed, 01 Nov 2023 07:07:57 +0100
+
 exim4 (4.96-15+deb12u2) bookworm-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-11-01 07:03:21.0 +0100
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actual EHLO LIMITS values with the
+ cached values and invalidate cache if different.  OK to carry on with
+ connect since values are advisory. */
+--

Bug#1049899: bookworm-pu: package exim4/4.96-15+deb12u3

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
Control: affects -1 + src:exim4

Hello,

I would like to push another round of cherry-picked upstream fixes to
bookworm, including the update to 4.96.2 to fix two non-DSA minor
security issues.

The changes are included in the new upstream (4.97 rc) uploads to sid which are 
present in sid and testing.


* Multiple bugfixes from upstream GIT master:
  + 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
  + 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
(Upstream bug 2998)
  + 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
  + 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
(Upstream bug 3013)
> ${run expansion breakage, similar to #1025420.
  + 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
TLS cert expiry date. Closes: #1043233
(Upstream bug 3014)
> This is major hickup, bordering on RC.

  + 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
> Another patch for ${run} expansion breakage.
  + 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
  + 76-12-DNS-more-hardening-against-crafted-responses.patch
* tests/basic: Add isolation-container restriction (needs a running
  exim daemon).
* Add ${run } expansion test to tests/basic.
* Update code to 4.96.2, fixing issues with the proxy protocol
  (CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42219). It
  also includes additional hardening for spf lookups, however CVE-2023-42218
  was diagnosed as a vulnerability in the libspf2 library and needs to be
  addressed there. Closes: #1053310

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-09-29 22:38:02.0 +0200
+++ exim4-4.96/debian/changelog	2023-11-01 07:07:57.0 +0100
@@ -1,3 +1,29 @@
+exim4 (4.96-15+deb12u3) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
++ 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
++ 76-10-Fix-tr.-and-empty-strings.-Bug-3023.patch ((Upstream bug 3023)
++ 76-12-DNS-more-hardening-against-crafted-responses.patch
+  * tests/basic: Add isolation-container restriction (needs a running
+exim daemon).
+  * Add ${run } expansion test to tests/basic.
+  * Update code to 4.96.2, fixing issues with the proxy protocol
+(CVE-2023-42117) and the `dnsdb` lookup subsystem (CVE-2023-42219). It
+also includes additional hardening for spf lookups, however CVE-2023-42218
+was diagnosed as a vulnerability in the libspf2 library and needs to be
+addressed there. Closes: #1053310
+
+ -- Andreas Metzler   Wed, 01 Nov 2023 07:07:57 +0100
+
 exim4 (4.96-15+deb12u2) bookworm-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-11-01 07:03:21.0 +0100
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actu

Bug#1049899: bookworm-pu: package exim4/4.96-15+deb12u2

[Andreas Metzler]

On 2023-10-07 Jonathan Wiltshire  wrote:
[...]
> The version number in this request matches one we've had via a DSA (5512);
> are they the same or does the proposed upload supercede it?
[...]


Hello,

I will need to rebase the proposed changes version on top of the DSA. I
got early notice that a security update was going to be needed and
expected that the timing had very good chance to conflict with the
stable update. (Which it did.) I therefore did not actually upload the
stable update. I only sent off-list notice about a delay (to Adam)
because the security issue was embargoed.

As of now there are still three open exim issues with too little info
but I still expect two more patches to exim and one for libspf. So I
will wait a little bit more before proposing another stable-upload. - Is
it alright to keep this bug open or should I close this one and reopen
another one when I am ready?

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#1049899: bookworm-pu: package exim4/4.96-15+deb12u2

[Andreas Metzler]

On 2023-08-16 Andreas Metzler  wrote:
[...]
> I would like to push another round of cherry-picked upstream fixes to
> bookworm. They have been part of the uploads to sid up to and including
> 4.96-19.
[...]

Hello,

I had to update the update since 75_78-Fix-free-of-value-after-run.patch
broke a specific expansion. While at it I also pulled the CI related
changes from -21.

cu Andreas

diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-07-02 14:56:17.0 +0200
+++ exim4-4.96/debian/changelog	2023-09-03 13:34:15.0 +0200
@@ -1,3 +1,22 @@
+exim4 (4.96-15+deb12u2) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
++ 75_83-Re-fix-live-variable-value-free.-The-inital-fix-resu.patch
+  * tests/basic: Add isolation-container restriction (needs a running
+exim daemon).
+  * Add ${run } expansion test to tests/basic.
+
+ -- Andreas Metzler   Sun, 03 Sep 2023 13:34:15 +0200
+
 exim4 (4.96-15+deb12u1) bookworm; urgency=medium
 
   * 75_42-Fix-run-arg-parsing.patch (From upstream GIT master, backported by
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-08-16 17:44:32.0 +0200
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actual EHLO LIMITS values with the
+ cached values and invalidate cache if different.  OK to carry on with
+ connect since values are advisory. */
+-- 
+2.40.1
+
diff -Nru exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
--- exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	2023-09-03 13:16:57.0 +0200
@@ -0,0 +1,99 @@
+From 1209e3e19e292cee517e43a2ccfe9b44b33bb1dc Mon Sep 17 00:00:00 2001
+From: Jasen Betts 
+Date: Sun, 23 Jul 2023 13:43:59 +0100
+Subject: [PATCH] Expansions: disallow UTF-16 surrogates from ${utf8clean:...}.
+  Bug 2998
+
+---
+ doc/ChangeLog |  4 
+ src/expand.c  | 27 +--
+ 2 files changed, 21 insertions(+), 10 deletions(-)
+
+--- a/src/expand.c
 b/src/expand.c
+@@ -7731,11 +7731,11 @@ NOT_ITEM: ;
+ 
+ 	case EOP_UTF8CLEAN:
+ 	  {
+ 	  int seq_len = 0, index = 0;
+ 	  int bytes_left = 0;
+-	  long codepoint = -1;
++	  ulong codepoint = (ulong)-1;
+ 	  int complete;
+ 	  uschar seq_buff[4];			/* accumulate utf-8 here */
+ 
+ 	  /* Manually track tainting, as we deal in individual chars below */
+ 
+@@ -7761,40 +7761,47 @@ NOT_ITEM: ;
+ 		codepoint = (codepoint << 6) | (c & 0x3f);
+ 		seq_buff[index++] = c;
+ 		if (--bytes_left == 0)		/* codepoint complete */
+ 		  if(codepoint > 0x10)	/* is it too large? */
+ 		complete = -1;	/* error (RFC3629 limit) */
++		  else if ( (codepoint & 0x1FF800 ) == 0xD800 ) /* surrogate */
++		/* A UTF-16 surrogate (which should be one of a pair that
++		encode a Unicode codepoint that is outside the Basic
++		Multilingual Plane).  Error, not UTF8.
++		RFC2279.2 is slight

Bug#1049899: bookworm-pu: package exim4/4.96-15+deb12u2

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
Control: affects -1 + src:exim4

Hello,

I would like to push another round of cherry-picked upstream fixes to
bookworm. They have been part of the uploads to sid up to and including
4.96-19.

This was mainly triggered by

+ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
  TLS cert expiry date. Closes: #1043233

which borders on RC and

+ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
  (Upstream bug 3013)

which is similar to #1025420.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-07-02 14:56:17.0 +0200
+++ exim4-4.96/debian/changelog	2023-08-16 15:12:39.0 +0200
@@ -1,3 +1,19 @@
+exim4 (4.96-15+deb12u2) bookworm; urgency=medium
+
+  * Multiple bugfixes from upstream GIT master:
++ 75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
++ 75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
+  (Upstream bug 2998)
++ 75_77-GnuTLS-fix-crash-with-tls_dhparam-none.patch
++ 75_78-Fix-free-of-value-after-run.patch
++ 75_79-Fix-recipients-expansion-when-used-within-run.-.-Bug.patch
+  (Upstream bug 3013)
++ 75_82-GnuTLS-fix-autogen-cert-expiry-date.-Bug-3014.patch: Fix on-demand
+  TLS cert expiry date. Closes: #1043233
+  (Upstream bug 3014)
+
+ -- Andreas Metzler   Wed, 16 Aug 2023 15:12:39 +0200
+
 exim4 (4.96-15+deb12u1) bookworm; urgency=medium
 
   * 75_42-Fix-run-arg-parsing.patch (From upstream GIT master, backported by
diff -Nru exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch
--- exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_74-Cancel-early-pipe-on-an-observed-advertising-change.patch	2023-08-16 14:59:06.0 +0200
@@ -0,0 +1,35 @@
+From 4d108ee9b8e5fb212c31812fef61529cd414 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Mon, 12 Jun 2023 22:13:46 +0100
+Subject: [PATCH] Cancel early-pipe on an observed advertising change
+
+---
+ src/transports/smtp.c | 3 +++
+ 1 file changed, 3 insertions(+)
+
+diff --git a/src/transports/smtp.c b/src/transports/smtp.c
+index c72028ce9..24ee577a2 100644
+--- a/src/transports/smtp.c
 b/src/transports/smtp.c
+@@ -,15 +,18 @@ if (pending_EHLO)
+   *(tls_out.active.sock < 0
+ 	? >ehlo_resp.cleartext_features : >ehlo_resp.crypted_features) =
+ 	  peer_offered;
+   *ap = authbits;
+   write_ehlo_cache_entry(sx);
+   }
+ else
++  {
+   invalidate_ehlo_cache_entry(sx);
++  sx->early_pipe_active = FALSE;	/* cancel further early-pipe on this conn */
++  }
+ 
+ return OK;		/* just carry on */
+ }
+ # ifdef EXPERIMENTAL_ESMTP_LIMITS
+ /* If we are handling LIMITS, compare the actual EHLO LIMITS values with the
+ cached values and invalidate cache if different.  OK to carry on with
+ connect since values are advisory. */
+-- 
+2.40.1
+
diff -Nru exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch
--- exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.96/debian/patches/75_76-Expansions-disallow-UTF-16-surrogates-from-utf8clean.patch	2023-08-16 15:11:54.0 +0200
@@ -0,0 +1,99 @@
+From 1209e3e19e292cee517e43a2ccfe9b44b33bb1dc Mon Sep 17 00:00:00 2001
+From: Jasen Betts 
+Date: Sun, 23 Jul 2023 13:43:59 +0100
+Subject: [PATCH] Expansions: disallow UTF-16 surrogates from ${utf8clean:...}.
+  Bug 2998
+
+---
+ doc/ChangeLog |  4 
+ src/expand.c  | 27 +--
+ 2 files changed, 21 insertions(+), 10 deletions(-)
+
+--- a/src/expand.c
 b/src/expand.c
+@@ -7731,11 +7731,11 @@ NOT_ITEM: ;
+ 
+ 	case EOP_UTF8CLEAN:
+ 	  {
+ 	  int seq_len = 0, index = 0;
+ 	  int bytes_left = 0;
+-	  long codepoint = -1;
++	  ulong codepoint = (ulong)-1;
+ 	  int complete;
+ 	  uschar seq_buff[4];			/* accumulate utf-8 here */
+ 
+ 	  /* Manually track tainting, as we deal in individual chars below */
+ 
+@@ -7761,40 +7761,47 @@ NOT_ITEM: ;
+ 		codepoint = (codepoint << 6) | (c & 0x3f);
+ 		seq_buff[index++] = c;
+ 		if (--bytes_left == 0)		/* codepoint complete */
+ 		  if(codepoint > 0x10)	/* is it too large? */
+ 		complete = -1;	/* error (RFC3629 limit) */
++		  else if ( (codepoint & 0x1FF800 ) == 0xD800 ) /* surrogate */
++		/* A UTF-16 surrogate (

Bug#1040139: bookworm-pu: package exim4/4.96-15

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: ex...@packages.debian.org
Control: affects -1 + src:exim4

Hello,

I would like to get most of the changes from 4.96-16 (unstable/testing)
into bookworm:
   * 75_42-Fix-run-arg-parsing.patch (From upstream GIT master, backported by
 Bryce Harrington for Ubuntu):  Fix argument parsing for ${run } expansion.
 Previously, when an argument included a close-brace character (eg. it
 itself used an expansion) an error occurred. Closes: #1025420
   * 75_68-Fix-srs_encode-.-for-mod-1024-day-zero.patch from upstream GIT
 master:  Fix ${srs_encode ..}. Previously it would give a bad result for
 one day every 1024 days.

The former is something has already popped up a couple of times on the
upstream user support mailing list.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff --git a/debian/changelog b/debian/changelog
index fbbb8c20..0231dc69 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,15 @@
+exim4 (4.96-15+deb12u1) bookworm; urgency=medium
+
+  * 75_42-Fix-run-arg-parsing.patch (From upstream GIT master, backported by
+Bryce Harrington for Ubuntu):  Fix argument parsing for ${run } expansion.
+Previously, when an argument included a close-brace character (eg. it
+itself used an expansion) an error occurred. Closes: #1025420
+  * 75_68-Fix-srs_encode-.-for-mod-1024-day-zero.patch from upstream GIT
+master:  Fix ${srs_encode ..}. Previously it would give a bad result for
+one day every 1024 days.
+
+ -- Andreas Metzler   Sun, 02 Jul 2023 14:56:17 +0200
+
 exim4 (4.96-15) unstable; urgency=medium
 
   * Pull from upstream GIT master:
diff --git a/debian/patches/75_42-Fix-run-arg-parsing.patch b/debian/patches/75_42-Fix-run-arg-parsing.patch
new file mode 100644
index ..79e55d61
--- /dev/null
+++ b/debian/patches/75_42-Fix-run-arg-parsing.patch
@@ -0,0 +1,100 @@
+From 44b6e099b76f403a55e77650821f8a69e9d2682e Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Sat, 3 Dec 2022 23:13:53 +
+Subject: [PATCH] Fix ${run } arg parsing
+ .
+  Backported by Bryce Harrington for Ubuntu
+
+Broken-by: cfe6acff2ddc
+---
+ doc/ChangeLog|  4 
+ src/expand.c | 13 ++---
+ src/transport.c  |  4 +++-
+ test/scripts/-Basic/0002 |  2 ++
+ test/stdout/0002 |  2 ++
+ 5 files changed, 21 insertions(+), 4 deletions(-)
+
+--- a/doc/ChangeLog
 b/doc/ChangeLog
+@@ -28,10 +28,14 @@ JH/13 Bug 2929: Fix using $recipients af
+ JH/14 Bug 2933: Fix regex substring match variables for null matches. Since 4.96
+   a capture group which obtained no text (eg. "(abc)*" matching zero
+   occurrences) could cause a segfault if the corresponding $ was
+   expanded.
+ 
++JH/15 Fix argument parsing for ${run } expansion. Previously, when an argument
++  included a close-brace character (eg. it itself used an expansion) an
++  error occurred.
++
+ 
+ 
+ Exim version 4.96
+ -
+ 
+--- a/src/expand.c
 b/src/expand.c
+@@ -5529,11 +5529,11 @@ while (*s)
+   {
+   FILE * f;
+   const uschar * arg, ** argv;
+   BOOL late_expand = TRUE;
+ 
+-  if ((expand_forbid & RDO_RUN) != 0)
++  if (expand_forbid & RDO_RUN)
+ {
+ expand_string_message = US"running a command is not permitted";
+ goto EXPAND_FAILED;
+ }
+ 
+@@ -5561,16 +5561,22 @@ while (*s)
+ 	}
+   s++;
+ 
+   if (late_expand)		/* this is the default case */
+ 	{
+-	int n = Ustrcspn(s, "}");
++	int n;
++	const uschar * t;
++	/* Locate the end of the args */
++	(void) expand_string_internal(s, TRUE, , TRUE, TRUE, NULL);
++	n = t - s;
+ 	arg = skipping ? NULL : string_copyn(s, n);
+ 	s += n;
+ 	}
+   else
+ 	{
++	DEBUG(D_expand)
++	  debug_printf_indent("args string for ${run} expand before split\n");
+ 	if (!(arg = expand_string_internal(s, TRUE, , skipping, TRUE, )))
+ 	  goto EXPAND_FAILED;
+ 	Uskip_whitespace();
+ 	}
+ 			/*{*/
+--- a/src/transport.c
 b/src/transport.c
+@@ -2187,10 +2187,12 @@ if (expand_arguments)
+   BOOL allow_dollar_recipients = addr && addr->parent
+ && Ustrcmp(addr->parent->address, "system-filter") == 0;
+ 
+   for (int i = 0; argv[i]; i++)
+ {
++DEBUG(D_expand) debug_printf_indent("arg %d\n", i);
++
+ /* Handle special fudge for passing an address list */
+ 
+ if (addr &&
+ (Ustrcmp(argv[i], "$pipe_addresses") == 0 ||
+  Ustrcmp(argv[i], "${pipe_addresses}") == 0))
+@@ -2361,11 +2363,11 @@ if (expand_arguments)
+   }
+ else *errptr = msg;
+ return FALSE;
+ }
+ 
+-  if ( f.running_in_test_harness && is_tainted(ex

Bug#1036395: RM: python-ooolib/0.0.22-5

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: rm
X-Debbugs-Cc: python-ooo...@packages.debian.org, ametz...@bebt.de
Control: affects -1 + src:python-ooolib

Please remove the package from testing, the maintainer was (silently)
waiting for it to autormed.
See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=965794#38

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#1036025: unblock: exim4/4.96-15

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: ex...@packages.debian.org
Control: affects -1 + src:exim4

Please unblock package exim4

This fixes an initialzation error which caused a crash in the smtp
transport. See https://bugs.exim.org/show_bug.cgi?id=2996

It is a one-line change, I have also removed garbage (unapplied patch)
from debian/patches.

cu Andreas

unblock exim4/4.96-15

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.96/debian/changelog exim4-4.96/debian/changelog
--- exim4-4.96/debian/changelog	2023-02-04 13:33:50.0 +0100
+++ exim4-4.96/debian/changelog	2023-05-10 18:30:35.0 +0200
@@ -1,3 +1,12 @@
+exim4 (4.96-15) unstable; urgency=medium
+
+  * Pull from upstream GIT master:
++ 75_70-Fix-variable-initialisation-in-smtp-transport.-Bug-2.patch
+  Fix a crash in the smtp transport.
+  https://bugs.exim.org/show_bug.cgi?id=2996
+
+ -- Andreas Metzler   Wed, 10 May 2023 18:30:35 +0200
+
 exim4 (4.96-14) unstable; urgency=medium
 
   * Pull from upstream GIT master:
diff -Nru exim4-4.96/debian/patches/75_42-Fix-run-arg-parsing.patch exim4-4.96/debian/patches/75_42-Fix-run-arg-parsing.patch
--- exim4-4.96/debian/patches/75_42-Fix-run-arg-parsing.patch	2022-12-04 08:02:50.0 +0100
+++ exim4-4.96/debian/patches/75_42-Fix-run-arg-parsing.patch	1970-01-01 01:00:00.0 +0100
@@ -1,99 +0,0 @@
-From 44b6e099b76f403a55e77650821f8a69e9d2682e Mon Sep 17 00:00:00 2001
-From: Jeremy Harris 
-Date: Sat, 3 Dec 2022 23:13:53 +
-Subject: [PATCH] Fix ${run } arg parsing
-
-Broken-by: cfe6acff2ddc

- doc/ChangeLog|  4 
- src/expand.c | 13 ++---
- src/transport.c  |  4 +++-
- test/scripts/-Basic/0002 |  2 ++
- test/stdout/0002 |  2 ++
- 5 files changed, 21 insertions(+), 4 deletions(-)
-
 a/doc/ChangeLog
-+++ b/doc/ChangeLog
-@@ -28,10 +28,14 @@
- JH/14 Bug 2933: Fix regex substring match variables for null matches. Since 4.96
-   a capture group which obtained no text (eg. "(abc)*" matching zero
-   occurrences) could cause a segfault if the corresponding $ was
-   expanded.
- 
-+JH/15 Fix argument parsing for ${run } expansion. Previously, when an argument
-+  included a close-brace character (eg. it itself used an expansion) an
-+  error occurred.
-+
- 
- 
- Exim version 4.96
- -
- 
 a/src/expand.c
-+++ b/src/expand.c
-@@ -5529,11 +5529,11 @@
-   {
-   FILE * f;
-   const uschar * arg, ** argv;
-   BOOL late_expand = TRUE;
- 
--  if ((expand_forbid & RDO_RUN) != 0)
-+  if (expand_forbid & RDO_RUN)
- {
- expand_string_message = US"running a command is not permitted";
- goto EXPAND_FAILED;
- }
- 
-@@ -5561,16 +5561,23 @@
- 	}
-   s++;
- 
-   if (late_expand)		/* this is the default case */
- 	{
--	int n = Ustrcspn(s, "}");
-+	int n;
-+	const uschar * t;
-+	/* Locate the end of the args */
-+	(void) expand_string_internal(s,
-+	  ESI_BRACE_ENDS | ESI_HONOR_DOLLAR | ESI_SKIPPING, , NULL, NULL);
-+	n = t - s;
- 	arg = skipping ? NULL : string_copyn(s, n);
- 	s += n;
- 	}
-   else
- 	{
-+	DEBUG(D_expand)
-+	  debug_printf_indent("args string for ${run} expand before split\n");
- 	if (!(arg = expand_string_internal(s, TRUE, , skipping, TRUE, )))
- 	  goto EXPAND_FAILED;
- 	Uskip_whitespace();
- 	}
- 			/*{*/
 a/src/transport.c
-+++ b/src/transport.c
-@@ -2187,10 +2187,12 @@
-   BOOL allow_dollar_recipients = addr && addr->parent
- && Ustrcmp(addr->parent->address, "system-filter") == 0;
- 
-   for (int i = 0; argv[i]; i++)
- {
-+DEBUG(D_expand) debug_printf_indent("arg %d\n", i);
-+
- /* Handle special fudge for passing an address list */
- 
- if (addr &&
- (Ustrcmp(argv[i], "$pipe_addresses") == 0 ||
-  Ustrcmp(argv[i], "${pipe_addresses}") == 0))
-@@ -2361,11 +2363,11 @@
-   }
- else *errptr = msg;
- return FALSE;
- }
- 
--  if ( f.running_in_test_harness && is_tainted(expanded_arg)
-+  if (  f.running_in_test_harness && is_tainted(expanded_arg)
- 	 && Ustrcmp(etext, "queryprogram router") == 0)
- 	{			/* hack, would be good to not need it */
- 	DEBUG(D_transport)
- 	  debug_printf("SPECIFIC TESTSUITE EXEMPTION: tainted arg '%s'\n",
- 		  expanded_arg);
diff -Nru exim4-4.96/debian/patches/75_70-Fix-variable-initialisation-in-smtp-transport.-Bug-2.patch exim4-4.96/debian/patches/75_70-Fix-variable-initialisation-in-smtp-transport.-Bug-2.patch
--- exim4-4.96/debian/patches/75_70-Fix-variable-initialisation-in-smtp-transport.-Bug-2.patch	1970-01-01 01:00:00.

Bug#1036005: unblock: pcp/6.0.3-1.1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: p...@packages.debian.org, ametz...@bebt.de
Control: affects -1 + src:pcp

Hello,

Please unblock package pcp, this just adds missing Replaces: for
upgrades from stable to testing.

The NMU was originally uploaded to the delayed queue but I later got
the go ahead from the package maintainer(s) to move to 0-day.

unblock pcp/6.0.3-1.1

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru pcp-6.0.3/debian/changelog pcp-6.0.3/debian/changelog
--- pcp-6.0.3/debian/changelog	2023-02-23 00:52:31.0 +0100
+++ pcp-6.0.3/debian/changelog	2023-05-07 11:36:37.0 +0200
@@ -1,3 +1,18 @@
+pcp (6.0.3-1.1) unstable; urgency=medium
+
+  * Non-maintainer upload.
+  * Add missing replaces on pcp (<< 5.3.0):
++ pcp-export-pcp2elasticsearch Closes: #1034998
++ pcp-export-pcp2json Closes: #1034983
++ pcp-export-pcp2spark Closes: #1034932
++ pcp-export-pcp2xml Closes: #1034922
++ pcp-export-pcp2zabbix Closes: #1034973
++ pcp-doc Closes: #1034966
++ pcp-zeroconf Closes: #1034895
++ pcp-export-pcp2xlsx (unfiled)
+
+ -- Andreas Metzler   Sun, 07 May 2023 11:36:37 +0200
+
 pcp (6.0.3-1) unstable; urgency=low
 
   * New release (full details in CHANGELOG).
diff -Nru pcp-6.0.3/debian/control pcp-6.0.3/debian/control
--- pcp-6.0.3/debian/control	2023-02-23 00:52:31.0 +0100
+++ pcp-6.0.3/debian/control	2023-05-07 07:37:15.0 +0200
@@ -385,6 +385,7 @@
 Package: pcp-export-pcp2elasticsearch
 Depends: python3-pcp, python3-requests, ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to Elasticsearch
  Performance Co-Pilot (PCP) front-end tool for exporting data from PCP
@@ -409,6 +410,7 @@
 Package: pcp-export-pcp2json
 Depends: python3-pcp, ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to JSON
  Performance Co-Pilot (PCP) front-end tool for exporting data from PCP
@@ -417,6 +419,7 @@
 Package: pcp-export-pcp2spark
 Depends: python3-pcp, python3-requests, ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to Apache Spark
  Performance Co-Pilot (PCP) front-end tools for exporting metric values
@@ -425,6 +428,7 @@
 Package: pcp-export-pcp2xml
 Depends: python3-pcp, ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to XML
  Performance Co-Pilot (PCP) front-end tool for exporting data from PCP
@@ -433,6 +437,7 @@
 Package: pcp-export-pcp2zabbix
 Depends: python3-pcp, ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to Zabbix
  Performance Co-Pilot (PCP) front-end tool for exporting data from PCP
@@ -449,7 +454,7 @@
 Section: doc
 Depends: ${misc:Depends}
 Breaks: pcp (<< 5.3.0), pcp-gui (<< 1.5.13)
-Replaces: pcp-gui (<< 1.5.13)
+Replaces: pcp (<< 5.3.0), pcp-gui (<< 1.5.13)
 Suggests: pcp, pcp-gui
 Architecture: all
 Description: Documentation and tutorial for the Performance Co-Pilot
@@ -462,6 +467,7 @@
 Package: pcp-zeroconf
 Depends: ${misc:Depends}, ${shlibs:Depends}, pcp (= ${binary:Version})
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Performance Co-Pilot (PCP) Zeroconf Package
  Contains configuration tweaks and files that increase metrics gathering
@@ -511,6 +517,7 @@
 Package: pcp-export-pcp2xlsx
 Depends: ${python:Depends}, ${misc:Depends}, ${shlibs:Depends}
 Breaks: pcp (<< 5.3.0)
+Replaces: pcp (<< 5.3.0)
 Architecture: any
 Description: Tool for exporting data from PCP to Excel spreadsheets
  Performance Co-Pilot (PCP) front-end tool for exporting data from PCP


signature.asc
Description: PGP signature

Re: SONAME bumps (transitions) always via experimental

[Andreas Metzler]

On 2023-01-10 Sam Hartman  wrote:
> > "Graham" == Graham Inggs  writes:
> Graham> Hi All
> Graham> On Fri, 6 Jan 2023 at 00:33, Bastian Blank  
> wrote:
> Graham> Would it be a bad thing to require all uploads that need to
> Graham> go through NEW (source and binary) to target experimental?

> Graham> I have been doing this for my own, and sponsored, uploads
> Graham> for some years already.

[...]
> Also, I'm less convinced there's a good reason for source new uploads to
> target experimental.
> If it's a new package with entirely new binary packages, it shouldn't be
> involved in any transitions.
[...]

Afaiui Graham's *question* was in response to Bastian's "However, please
describe an actionable plan." Obviously it would be a lot easier if we
could require to have all NEW uploads go to experimental instead of
trying to filter for soname bumps.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Re: Understanding what is blocking spamassassin 4.0.0 testing migration

[Andreas Metzler]

On 2022-12-29 "Adam D. Barratt"  wrote:
> On Thu, 2022-12-29 at 07:21 +0100, Andreas Metzler wrote:
[...]
> > removing spamassassin/4.0.0~rc4-1/amd64 from testing makes claws-
[...]

> That's due to the arch:all build failing, which means there is no
> "spamassassin" binary package in unstable currently when combined with
> dak's feature of hiding arch:all packages that don't correspond to the
> version of arch-dep packages. See #915948 for further details on
> similar issues.

> Looking at previous build failures, it looks like the arch:all build
> has issues with IPv6-only buildds. I've given it back, so let's see
> what happens.

Thank you (and Paul) for the explanation. I did not get that "removing
spamassassin" was talking about the binary. The retried build succeeded.
:-)

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Understanding what is blocking spamassassin 4.0.0 testing migration

[Andreas Metzler]

Hello,

I do not understand why spamassassin 4.0.0 does not prpagate to testing.
Tracker/excuses https://qa.debian.org/excuses.php?package=spamassassin
says:
Issues preventing migration:
[...]
removing spamassassin/4.0.0~rc4-1/amd64 from testing makes 
claws-mail-spamassassin/4.1.1-2/amd64 uninstallable
removing spamassassin/4.0.0~rc4-1/amd64 from testing makes 
evolution-plugin-spamassassin/3.46.2-1/amd64 uninstallable
[ list of more spamassassin rdeps which would be uninstallable if ]


Note that it only talks about "removing" instead of "upgrading". Which
obviously cannot work.

https://release.debian.org/britney/update_output.txt does not mention
spamassassin at all. It also seems very short, not like a full run with
less than 1000 lines.

TIA, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#1023261: bullseye-pu: package libtasn1-6/4.16.0-2+deb11u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bullseye
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: libtasn...@packages.debian.org, t...@security.debian.org

Hello,

I would like to fix CVE-2021-46848 in bullseye. This was fixed in
sid/testing by new upstream 4.19.0. I already had some correspondence
with debian-security, no DSA is planned.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru libtasn1-6-4.16.0/debian/changelog libtasn1-6-4.16.0/debian/changelog
--- libtasn1-6-4.16.0/debian/changelog	2020-02-15 17:38:59.0 +0100
+++ libtasn1-6-4.16.0/debian/changelog	2022-11-01 11:57:42.0 +0100
@@ -1,3 +1,10 @@
+libtasn1-6 (4.16.0-2+deb11u1) bullseye; urgency=medium
+
+  * Fix ETYPE_OK out of bounds read. CVE-2021-46848
+10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch
+
+ -- Andreas Metzler   Tue, 01 Nov 2022 11:57:42 +0100
+
 libtasn1-6 (4.16.0-2) unstable; urgency=low
 
   * Upload to unstable.
diff -Nru libtasn1-6-4.16.0/debian/patches/10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch libtasn1-6-4.16.0/debian/patches/10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch
--- libtasn1-6-4.16.0/debian/patches/10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch	1970-01-01 01:00:00.0 +0100
+++ libtasn1-6-4.16.0/debian/patches/10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch	2022-10-30 13:02:08.0 +0100
@@ -0,0 +1,29 @@
+From 44a700d2051a666235748970c2df047ff207aeb5 Mon Sep 17 00:00:00 2001
+From: Simon Josefsson 
+Date: Wed, 17 Aug 2022 12:25:06 +0200
+Subject: [PATCH] Fix ETYPE_OK off by one array size check.  Closes: #32.
+
+Reported by David Trabish in
+<https://gitlab.com/gnutls/libtasn1/-/issues/32>.
+
+Signed-off-by: Simon Josefsson 
+---
+ NEWS  | 1 +
+ lib/int.h | 2 +-
+ 2 files changed, 2 insertions(+), 1 deletion(-)
+
+--- a/lib/int.h
 b/lib/int.h
+@@ -95,11 +95,11 @@
+ 	case ASN1_ETYPE_SET_OF
+ 
+ #define ETYPE_TAG(etype) (_asn1_tags[etype].tag)
+ #define ETYPE_CLASS(etype) (_asn1_tags[etype].class)
+ #define ETYPE_OK(etype) (((etype) != ASN1_ETYPE_INVALID && \
+-  (etype) <= _asn1_tags_size && \
++  (etype) < _asn1_tags_size && \
+   _asn1_tags[(etype)].desc != NULL)?1:0)
+ 
+ #define ETYPE_IS_STRING(etype) ((etype == ASN1_ETYPE_GENERALSTRING || \
+ 	etype == ASN1_ETYPE_NUMERIC_STRING || etype == ASN1_ETYPE_IA5_STRING || \
+ 	etype == ASN1_ETYPE_TELETEX_STRING || etype == ASN1_ETYPE_PRINTABLE_STRING || \
diff -Nru libtasn1-6-4.16.0/debian/patches/series libtasn1-6-4.16.0/debian/patches/series
--- libtasn1-6-4.16.0/debian/patches/series	1970-01-01 01:00:00.0 +0100
+++ libtasn1-6-4.16.0/debian/patches/series	2022-11-01 11:57:42.0 +0100
@@ -0,0 +1 @@
+10_Fix-ETYPE_OK-off-by-one-array-size-check.-Closes-32.patch


signature.asc
Description: PGP signature

Bug#1019876: nmu: atlas-ecmwf_0.30.0-3

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: binnmu
X-Debbugs-Cc: ametz...@bebt.de, dh-fortran-...@packages.debian.org

Hello,

according to 
https://binarycontrol.debian.net/?q=%5Ermdir+--ignore-fail-on-non-empty+%2Fusr%2Flib%2F%5C%24multiarch%2Ffortran%2Fgfortran=
there are still 4 packages built with broken dh-fortran-mod.

nmu atlas-ecmwf_0.30.0-3 . ANY . unstable . -m "Rebuild with fixed 
dh-fortran-mod (See #1019050)"
nmu cdo_2.0.6-2 . ANY . unstable . -m "Rebuild with fixed dh-fortran-mod (See 
#1019050)"
nmu petsc_3.17.4+dfsg1-1 . ANY . unstable . -m "Rebuild with fixed 
dh-fortran-mod (See #1019050)"
nmu slepc_3.17.2+dfsg1-2 . ANY . unstable . -m "Rebuild with fixed 
dh-fortran-mod (See #1019050)"

TIA, cu Andreas

Bug#1014474: nmu: sa-exim_4.2.1-20

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: binnmu

nmu sa-exim_4.2.1-20 . ANY . unstable . -m "rebuild against exim4 4.96"

Please rebuild sa-exim allowing exim4 4.96-1 to propagate to testing.

TIA, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#1012033: bullseye-pu: package gnutls28/3.7.1-5+deb11u1

[Andreas Metzler]

On 2022-05-29 Andreas Metzler  wrote:
[...]
> as requested in #1011246 I would like fix miscalculation of SHA384 in
> the SSA accelarated implementation.

> It is a one-line change and was part of the 3.7.3 release.
[...]

Actually this seems like a good opportunity to fix a minor CVE, which
was also fixed in 3.7.3.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.7.1/debian/changelog gnutls28-3.7.1/debian/changelog
--- gnutls28-3.7.1/debian/changelog 2021-05-29 12:14:30.0 +0200
+++ gnutls28-3.7.1/debian/changelog 2022-06-14 18:55:44.0 +0200
@@ -1,3 +1,12 @@
+gnutls28 (3.7.1-5+deb11u1) bullseye; urgency=medium
+
+  * 56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch: Backport SSSE3 SHA384
+miscalculation fix from 3.7.3.  Closes: #1011246
+  * 56_45-wrap_nettle_hash_fast-avoid-calling-_update-with-zer.patch from
+3.7.3: Fix null-pointer dereference flaw. CVE-2021-4209
+
+ -- Andreas Metzler   Tue, 14 Jun 2022 18:55:44 +0200
+
 gnutls28 (3.7.1-5) unstable; urgency=medium
 
   * Another fix from 3.7.2:
diff -Nru 
gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch
 
gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch
--- 
gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch
   1970-01-01 01:00:00.0 +0100
+++ 
gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch
   2022-06-14 10:55:13.0 +0200
@@ -0,0 +1,34 @@
+From acdfeb4b3f0c64ad20f28513618e9903bfb81426 Mon Sep 17 00:00:00 2001
+From: Miroslav Lichvar 
+Date: Wed, 1 Sep 2021 15:48:27 +0200
+Subject: [PATCH] fix SSSE3 SHA384 to work more than once
+
+The output function called sha512_digest() instead of sha384_digest(),
+which caused the hash context to be reinitialized for SHA512 instead of
+SHA384 and all following digests using the hash handle were wrong.
+
+Signed-off-by: Miroslav Lichvar 
+---
+ lib/accelerated/x86/sha-x86-ssse3.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/lib/accelerated/x86/sha-x86-ssse3.c 
b/lib/accelerated/x86/sha-x86-ssse3.c
+index 8ea4e54aee..1d442e97e7 100644
+--- a/lib/accelerated/x86/sha-x86-ssse3.c
 b/lib/accelerated/x86/sha-x86-ssse3.c
+@@ -258,11 +258,11 @@ static int _ctx_init(gnutls_digest_algorithm_t algo,
+   ctx->length = SHA256_DIGEST_SIZE;
+   break;
+   case GNUTLS_DIG_SHA384:
+   sha384_init(>ctx.sha384);
+   ctx->update = (update_func) x86_sha512_update;
+-  ctx->digest = (digest_func) sha512_digest;
++  ctx->digest = (digest_func) sha384_digest;
+   ctx->init = (init_func) sha384_init;
+   ctx->ctx_ptr = >ctx.sha384;
+   ctx->length = SHA384_DIGEST_SIZE;
+   break;
+   case GNUTLS_DIG_SHA512:
+-- 
+2.35.1
+
diff -Nru 
gnutls28-3.7.1/debian/patches/56_45-wrap_nettle_hash_fast-avoid-calling-_update-with-zer.patch
 
gnutls28-3.7.1/debian/patches/56_45-wrap_nettle_hash_fast-avoid-calling-_update-with-zer.patch
--- 
gnutls28-3.7.1/debian/patches/56_45-wrap_nettle_hash_fast-avoid-calling-_update-with-zer.patch
  1970-01-01 01:00:00.0 +0100
+++ 
gnutls28-3.7.1/debian/patches/56_45-wrap_nettle_hash_fast-avoid-calling-_update-with-zer.patch
  2022-06-14 10:58:46.0 +0200
@@ -0,0 +1,32 @@
+From 3db352734472d851318944db13be73da61300568 Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Wed, 22 Dec 2021 09:12:25 +0100
+Subject: [PATCH] wrap_nettle_hash_fast: avoid calling _update with zero-length
+ input
+
+As Nettle's hash update functions internally call memcpy, providing
+zero-length input may cause undefined behavior.
+
+Signed-off-by: Daiki Ueno 
+---
+ lib/nettle/mac.c | 4 +++-
+ 1 file changed, 3 insertions(+), 1 deletion(-)
+
+diff --git a/lib/nettle/mac.c b/lib/nettle/mac.c
+index f9d4d7a8df..35e070fab0 100644
+--- a/lib/nettle/mac.c
 b/lib/nettle/mac.c
+@@ -788,7 +788,9 @@ static int wrap_nettle_hash_fast(gnutls_digest_algorithm_t 
algo,
+   if (ret < 0)
+   return gnutls_assert_val(ret);
+ 
+-  ctx.update(, text_size, text);
++  if (text_size > 0) {
++  ctx.update(, text_size, text);
++  }
+   ctx.digest(, ctx.length, digest);
+ 
+   return 0;
+-- 
+2.35.1
+
diff -Nru gnutls28-3.7.1/debian/patches/series 
gnutls28-3.7.1/debian/patches/series
--- gnutls28-3.7.1/debian/patches/series2021-05-29 11:37:38.0 
+0200
+++ gnutls28-3.7.1/debian/patches/series2022-06-14 10:59:12.0 
+0200
@@ -18,3 +18,5 @@
 56_28-handshake-fix-timing-of-sending-early-data.patch
 56_30-x509-verify-treat-SHA-1-signed-CA-in-the-trusted-set.patch
 56_33-serv-stop-setting-AI_ADDRCONFIG-on-getaddrinfo.patch
+56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch

Bug#1012033: bullseye-pu: package gnutls28/3.7.1-5+deb11u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: bullseye
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: Dmitry Baryshkov , 
gnutl...@packages.debian.org

Hello,

as requested in #1011246 I would like fix miscalculation of SHA384 in
the SSA accelarated implementation.

It is a one-line change and was part of the 3.7.3 release.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.7.1/debian/changelog gnutls28-3.7.1/debian/changelog
--- gnutls28-3.7.1/debian/changelog	2021-05-29 12:14:30.0 +0200
+++ gnutls28-3.7.1/debian/changelog	2022-05-22 13:04:01.0 +0200
@@ -1,3 +1,10 @@
+gnutls28 (3.7.1-5+deb11u1) bullseye; urgency=medium
+
+  * 56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch: Backport SSSE3 SHA384
+miscalculation fix from 3.7.3.  Closes: #1011246
+
+ -- Andreas Metzler   Sun, 22 May 2022 13:04:01 +0200
+
 gnutls28 (3.7.1-5) unstable; urgency=medium
 
   * Another fix from 3.7.2:
diff -Nru gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch
--- gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.7.1/debian/patches/56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch	2022-05-22 13:04:01.0 +0200
@@ -0,0 +1,34 @@
+From acdfeb4b3f0c64ad20f28513618e9903bfb81426 Mon Sep 17 00:00:00 2001
+From: Miroslav Lichvar 
+Date: Wed, 1 Sep 2021 15:48:27 +0200
+Subject: [PATCH] fix SSSE3 SHA384 to work more than once
+
+The output function called sha512_digest() instead of sha384_digest(),
+which caused the hash context to be reinitialized for SHA512 instead of
+SHA384 and all following digests using the hash handle were wrong.
+
+Signed-off-by: Miroslav Lichvar 
+---
+ lib/accelerated/x86/sha-x86-ssse3.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/lib/accelerated/x86/sha-x86-ssse3.c b/lib/accelerated/x86/sha-x86-ssse3.c
+index 8ea4e54aee..1d442e97e7 100644
+--- a/lib/accelerated/x86/sha-x86-ssse3.c
 b/lib/accelerated/x86/sha-x86-ssse3.c
+@@ -258,11 +258,11 @@ static int _ctx_init(gnutls_digest_algorithm_t algo,
+ 		ctx->length = SHA256_DIGEST_SIZE;
+ 		break;
+ 	case GNUTLS_DIG_SHA384:
+ 		sha384_init(>ctx.sha384);
+ 		ctx->update = (update_func) x86_sha512_update;
+-		ctx->digest = (digest_func) sha512_digest;
++		ctx->digest = (digest_func) sha384_digest;
+ 		ctx->init = (init_func) sha384_init;
+ 		ctx->ctx_ptr = >ctx.sha384;
+ 		ctx->length = SHA384_DIGEST_SIZE;
+ 		break;
+ 	case GNUTLS_DIG_SHA512:
+-- 
+2.35.1
+
diff -Nru gnutls28-3.7.1/debian/patches/series gnutls28-3.7.1/debian/patches/series
--- gnutls28-3.7.1/debian/patches/series	2021-05-29 11:37:38.0 +0200
+++ gnutls28-3.7.1/debian/patches/series	2022-05-22 13:04:01.0 +0200
@@ -18,3 +18,4 @@
 56_28-handshake-fix-timing-of-sending-early-data.patch
 56_30-x509-verify-treat-SHA-1-signed-CA-in-the-trusted-set.patch
 56_33-serv-stop-setting-AI_ADDRCONFIG-on-getaddrinfo.patch
+56_40-fix-SSSE3-SHA384-to-work-more-than-once.patch


signature.asc
Description: PGP signature

Bug#959469: openssl 1.1.1n-0+deb10u1 flagged for acceptance

[Andreas Metzler]

X-Debbugs-Cc: gnutl...@packages.debian.org, Kurt Roeckx , Paul 
Gevers , Sebastian Andrzej Siewior 

On 2022-03-21 Sebastian Andrzej Siewior  wrote:
> On 2022-03-21 00:12:11 [+0100], To Kurt Roeckx wrote:
> > doesn't help here but
> >  -cipher "ALL:@SECLEVEL=1"

> > does. 

> Only debci is affected. The package builds because this testsuite is not
> part of the build process.
> I prepared a NMU against Buster for gnutls. I can open later today a
> buster-pu and do the upload unless someone objects or gnutls folks have
> something in their queue.
> Please let me know.

Hello Sebastian,

thanks for taking care, feel free to NMU.

cu Andreas

PS: Style nitpick: As can be see from "ls debian/patches/" I think it is
a very good idea to use patch filenames with obvious sorting.


signature.asc
Description: PGP signature

Bug#991397: unblock: exim4/4.94.2-7

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: ex...@packages.debian.org, Adrian Bunk 

Please unblock package exim4

This is release fixes a single bug by pulling the respective fix from
upstream's +fixes branch.

When control=fakereject is used with a custom error message the
respective non-safe data was expanded. With allow_insecure_tainted_data
not set this only causes a entry in paniclog, otherwise the actual
expansion might happen.

Debian's default exim configuration does not use control=fakereject but
still I would consider this an important bug that I would like to see
fixed.

unblock exim4/4.94.2-7

Thanks, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.94.2/debian/changelog exim4-4.94.2/debian/changelog
--- exim4-4.94.2/debian/changelog	2021-05-26 18:49:44.0 +0200
+++ exim4-4.94.2/debian/changelog	2021-07-13 18:04:57.0 +0200
@@ -1,3 +1,10 @@
+exim4 (4.94.2-7) unstable; urgency=medium
+
+  * 73_05-Fix-tainted-message-for-fakereject.patch from upstream +fixes
+branch: Fix re-expansion of custom message with control=fakereject.
+
+ -- Andreas Metzler   Tue, 13 Jul 2021 18:04:57 +0200
+
 exim4 (4.94.2-6) unstable; urgency=medium
 
   * Cherrypick
diff -Nru exim4-4.94.2/debian/patches/73_05-Fix-tainted-message-for-fakereject.patch exim4-4.94.2/debian/patches/73_05-Fix-tainted-message-for-fakereject.patch
--- exim4-4.94.2/debian/patches/73_05-Fix-tainted-message-for-fakereject.patch	1970-01-01 01:00:00.0 +0100
+++ exim4-4.94.2/debian/patches/73_05-Fix-tainted-message-for-fakereject.patch	2021-07-13 18:03:04.0 +0200
@@ -0,0 +1,44 @@
+From c819f3bcad02bcb06004ae2ad135b68fab0ae888 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Wed, 7 Jul 2021 22:19:07 +0100
+Subject: [PATCH 5/5] Fix tainted message for fakereject
+
+(cherry picked from commit a9ac2d7fc219e41a353abf1f599258b9b9d21b7e)
+---
+ doc/ChangeLog | 4 
+ src/acl.c | 4 +++-
+ 2 files changed, 7 insertions(+), 1 deletion(-)
+
+diff --git a/doc/ChangeLog b/doc/ChangeLog
+index e60c1cad5..3e93f653f 100644
+--- a/doc/ChangeLog
 b/doc/ChangeLog
+@@ -227,6 +227,10 @@ JH/53 Bug 2743: fix immediate-delivery via named queue.  Previously this would
+   fail with a taint-check on the spoolfile name, and leave the message
+   queued.
+ 
++JH/57 Fix control=fakreject for a custom message containing tainted data.
++  Previously this resulted in a log complaint, due to a re-expansion present
++  since fakereject was originally introduced.
++
+ 
+ Exim version 4.94
+ -
+diff --git a/src/acl.c b/src/acl.c
+index 7061230b4..65324405c 100644
+--- a/src/acl.c
 b/src/acl.c
+@@ -3137,7 +3137,9 @@ for (; cb; cb = cb->next)
+ 	{
+ 	const uschar *pp = p + 1;
+ 	while (*pp) pp++;
+-	fake_response_text = expand_string(string_copyn(p+1, pp-p-1));
++	/* The entire control= line was expanded at top so no need to expand
++	the part after the / */
++	fake_response_text = string_copyn(p+1, pp-p-1);
+ 	p = pp;
+ 	}
+ 	   else /* Explicitly reset to default string */
+-- 
+2.30.2
+
diff -Nru exim4-4.94.2/debian/patches/series exim4-4.94.2/debian/patches/series
--- exim4-4.94.2/debian/patches/series	2021-05-22 13:27:33.0 +0200
+++ exim4-4.94.2/debian/patches/series	2021-07-13 18:03:23.0 +0200
@@ -10,6 +10,7 @@
 73_02-Fix-ipv6norm.patch
 73_03-Named-Queues-fix-immediate-delivery.-Bug-2743.patch
 73_04-Fix-host_name_lookup-Close-2747.patch
+73_05-Fix-tainted-message-for-fakereject.patch
 75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
 75_02-search.patch
 75_03-dbstuff.patch


signature.asc
Description: PGP signature

Bug#990919: unblock: exim4/4.94.2-6

[Andreas Metzler]

On 2021-07-11 Adrian Bunk  wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: unblock

> Please unblock package exim4

>   * Cherrypick
> 78_01-Command-line-option-for-no-notifier-socket.-Bug-2616.patch from
> upstream GIT master. This allows one to disable creation of a
> daemon notifier socket by either setting notifier_socket to a empty value
> or specifying -oY commandline option.
>   * Init script: For QUEUERUNNER='separate' start daemons with -oY commandline
> option to disable daemon notifier socket. Enforce lockstep ugrade of -base
> and *daemon* by temporarily adding a versioned Breaks to exim4-base on
> older *daemon*. Closes: #988844
> (change by Andreas Metzler)

> This fixes a regression from buster.

> Maintainer and bug submitter are in Cc, ack/nak would be appreciated.
[...]

Hello,

Thank you Adrian, this upload is indeed targeted for bullseyw but I
forgot to submit the unblock request.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#989422: buster-pu: package libgcrypt20/1.8.4-5+deb10u1

[Andreas Metzler]

On 2021-06-12 "Adam D. Barratt"  wrote:
[...]
> As we're getting close to the window for 10.10 closing, please feel
> free to upload the package and we'll handle the d-i coordination from
> there.

Thanks for the heads-up. Done.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#989422: buster-pu: package libgcrypt20/1.8.4-5+deb10u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: libgcryp...@packages.debian.org,secur...@debian.org


Hello,

I would like to fix the non-DSA CVE-2021-33560 for buster by
cherrypicking the respective commit from 1.8.8. This is about weak
ElGamal encyption when a key not generated by libgcrypt/gnupg is used.

This was fixed in unstable's 1.8.7-6, with bullseye unblock request
#989421 sent a couple of minutes ago.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru libgcrypt20-1.8.4/debian/changelog libgcrypt20-1.8.4/debian/changelog
--- libgcrypt20-1.8.4/debian/changelog	2019-01-20 14:47:23.0 +0100
+++ libgcrypt20-1.8.4/debian/changelog	2021-05-29 13:32:02.0 +0200
@@ -1,3 +1,11 @@
+libgcrypt20 (1.8.4-5+deb10u1) buster; urgency=medium
+
+  * 31_cipher-Fix-ElGamal-encryption-for-other-implementati.patch from
+upstream LIBGCRYPT-1.8-BRANCH: Fix weak ElGamal encryption with keys *not*
+generated by GnuPG/libgcrypt. CVE-2021-33560
+
+ -- Andreas Metzler   Sat, 29 May 2021 13:32:02 +0200
+
 libgcrypt20 (1.8.4-5) unstable; urgency=medium
 
   * 30_doc-Fix-library-initialization-examples.patch from upstream
diff -Nru libgcrypt20-1.8.4/debian/patches/31_cipher-Fix-ElGamal-encryption-for-other-implementati.patch libgcrypt20-1.8.4/debian/patches/31_cipher-Fix-ElGamal-encryption-for-other-implementati.patch
--- libgcrypt20-1.8.4/debian/patches/31_cipher-Fix-ElGamal-encryption-for-other-implementati.patch	1970-01-01 01:00:00.0 +0100
+++ libgcrypt20-1.8.4/debian/patches/31_cipher-Fix-ElGamal-encryption-for-other-implementati.patch	2021-05-29 13:16:14.0 +0200
@@ -0,0 +1,105 @@
+From 707c3c5c511ee70ad0e39ec613471f665305fbea Mon Sep 17 00:00:00 2001
+From: NIIBE Yutaka 
+Date: Fri, 21 May 2021 11:15:07 +0900
+Subject: [PATCH] cipher: Fix ElGamal encryption for other implementations.
+
+* cipher/elgamal.c (gen_k): Remove support of smaller K.
+(do_encrypt): Never use smaller K.
+(sign): Folllow the change of gen_k.
+
+--
+
+Cherry-pick master commit of:
+	632d80ef30e13de6926d503aa697f92b5dbfbc5e
+
+This change basically reverts encryption changes in two commits:
+
+	74386120dad6b3da62db37f7044267c8ef34689b
+	78531373a342aeb847950f404343a05e36022065
+
+Use of smaller K for ephemeral key in ElGamal encryption is only good,
+when we can guarantee that recipient's key is generated by our
+implementation (or compatible).
+
+For detail, please see:
+
+Luca De Feo, Bertram Poettering, Alessandro Sorniotti,
+"On the (in)security of ElGamal in OpenPGP";
+in the proceedings of  CCS'2021.
+
+CVE-id: CVE-2021-33560
+GnuPG-bug-id: 5328
+Suggested-by: Luca De Feo, Bertram Poettering, Alessandro Sorniotti
+Signed-off-by: NIIBE Yutaka 
+---
+ cipher/elgamal.c | 24 ++--
+ 1 file changed, 6 insertions(+), 18 deletions(-)
+
+diff --git a/cipher/elgamal.c b/cipher/elgamal.c
+index 4eb52d62..ae7a631e 100644
+--- a/cipher/elgamal.c
 b/cipher/elgamal.c
+@@ -66,7 +66,7 @@ static const char *elg_names[] =
+ 
+ 
+ static int test_keys (ELG_secret_key *sk, unsigned int nbits, int nodie);
+-static gcry_mpi_t gen_k (gcry_mpi_t p, int small_k);
++static gcry_mpi_t gen_k (gcry_mpi_t p);
+ static gcry_err_code_t generate (ELG_secret_key *sk, unsigned nbits,
+  gcry_mpi_t **factors);
+ static int  check_secret_key (ELG_secret_key *sk);
+@@ -189,11 +189,10 @@ test_keys ( ELG_secret_key *sk, unsigned int nbits, int nodie )
+ 
+ /
+  * Generate a random secret exponent k from prime p, so that k is
+- * relatively prime to p-1.  With SMALL_K set, k will be selected for
+- * better encryption performance - this must never be used signing!
++ * relatively prime to p-1.
+  */
+ static gcry_mpi_t
+-gen_k( gcry_mpi_t p, int small_k )
++gen_k( gcry_mpi_t p )
+ {
+   gcry_mpi_t k = mpi_alloc_secure( 0 );
+   gcry_mpi_t temp = mpi_alloc( mpi_get_nlimbs(p) );
+@@ -202,18 +201,7 @@ gen_k( gcry_mpi_t p, int small_k )
+   unsigned int nbits, nbytes;
+   char *rndbuf = NULL;
+ 
+-  if (small_k)
+-{
+-  /* Using a k much lesser than p is sufficient for encryption and
+-   * it greatly improves the encryption performance.  We use
+-   * Wiener's table and add a large safety margin. */
+-  nbits = wiener_map( orig_nbits ) * 3 / 2;
+-  if( nbits >= orig_nbits )
+-BUG();
+-}
+-  else
+-nbits = orig_nbits;
+-
++  nbits = orig_nbits;
+ 
+   nbytes = (nbits+7)/8;
+   if( DBG_CIPHER )
+@@ -492,7 +480,7 @@ do_encrypt(gcry_mpi_t a, gcry_mpi_t b, gcry_mpi_t input, ELG_public_key *pkey )
+* error code.
+*/
+ 
+-  k = gen_k( pkey->p, 1 );
++  k = gen_k( pkey->p );
+   mpi_powm (a, pkey->g, k, pkey->p);
+ 
+   /* b = (y^k * input) mod p
+@@ -594,7 +582,7 @@ sign(gcry_mpi_t a, gcry_mpi_t b, gcry_mpi_t in

Bug#989421: unblock: libgcrypt20/1.8.7-6

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: libgcryp...@packages.debian.org

Please unblock package libgcrypt20.

Compared to 1.8.7-3 this pulls a 4 commits from 1.8.8, including
30_10-cipher-Fix-ElGamal-encryption-for-other-implementati.patch
(CVE-2021-33560) which fixes weak ElGamal encryption with keys *not*
generated by libgcrypt/gnupg. It does not warrant a DSA (already
doublechecked with debian-security) but should still be fixed. I will
also prepare an upload for buster.

unblock libgcrypt20/1.8.7-6

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru libgcrypt20-1.8.7/debian/changelog libgcrypt20-1.8.7/debian/changelog
--- libgcrypt20-1.8.7/debian/changelog	2021-02-14 15:27:13.0 +0100
+++ libgcrypt20-1.8.7/debian/changelog	2021-05-27 18:07:38.0 +0200
@@ -1,3 +1,26 @@
+libgcrypt20 (1.8.7-6) unstable; urgency=medium
+
+  * Update from LIBGCRYPT-1.8-BRANCH:
++ 30_10-cipher-Fix-ElGamal-encryption-for-other-implementati.patch
+
+ -- Andreas Metzler   Thu, 27 May 2021 18:07:38 +0200
+
+libgcrypt20 (1.8.7-5) unstable; urgency=medium
+
+  * Pull fix for ECC decyryption regression (caused by
+30_08-ecc-Check-the-input-length-for-the-point.patch) from
+LIBGCRYPT-1.8-BRANCH. Closes: #987956
+
+ -- Andreas Metzler   Thu, 06 May 2021 18:06:14 +0200
+
+libgcrypt20 (1.8.7-4) unstable; urgency=medium
+
+  * Update from LIBGCRYPT-1.8-BRANCH:
++ 30_07-Fix-previous-commit.patch
++ 30_08-ecc-Check-the-input-length-for-the-point.patch
+
+ -- Andreas Metzler   Sun, 02 May 2021 13:58:47 +0200
+
 libgcrypt20 (1.8.7-3) unstable; urgency=medium
 
   * Update from LIBGCRYPT-1.8-BRANCH:
diff -Nru libgcrypt20-1.8.7/debian/patches/30_07-Fix-previous-commit.patch libgcrypt20-1.8.7/debian/patches/30_07-Fix-previous-commit.patch
--- libgcrypt20-1.8.7/debian/patches/30_07-Fix-previous-commit.patch	1970-01-01 01:00:00.0 +0100
+++ libgcrypt20-1.8.7/debian/patches/30_07-Fix-previous-commit.patch	2021-05-02 13:52:17.0 +0200
@@ -0,0 +1,41 @@
+From a5799f1618aaf1bbb52e7e121275228dd4a3ac8b Mon Sep 17 00:00:00 2001
+From: Werner Koch 
+Date: Sun, 14 Feb 2021 18:54:40 +0100
+Subject: [PATCH 7/8] Fix previous commit
+
+* src/global.c (_gcry_get_config): Append the Nul only in the !what
+case.
+--
+
+Fixes-commit: 3f42f727a0699f7274a99ea39def7f9b4c3b0c1e
+Actually this was my fault - I stripped off the test which Jussi did in
+his original fix on master.  And did not run make check.
+
+Signed-off-by: Werner Koch 
+---
+ src/global.c | 9 +++--
+ 1 file changed, 7 insertions(+), 2 deletions(-)
+
+diff --git a/src/global.c b/src/global.c
+index 7d634095..95daedac 100644
+--- a/src/global.c
 b/src/global.c
+@@ -419,8 +419,13 @@ _gcry_get_config (int mode, const char *what)
+ 
+   print_config (what, fp);
+ 
+-  /* Make sure the output is null terminated. */
+-  gpgrt_fwrite ("", 1, 1, fp);
++  /* Make sure the output is null terminated if no specific item was
++   * requested.  This is needed because tests/version.c expects that
++   * the function fails with the !data case below.  For the specific
++   * test an extra nul is not required because we always have a LF
++   * which is then replaced right at the end of this function.  */
++  if (!what)
++gpgrt_fwrite ("", 1, 1, fp);
+ 
+   if (gpgrt_ferror (fp))
+ {
+-- 
+2.30.2
+
diff -Nru libgcrypt20-1.8.7/debian/patches/30_08-ecc-Check-the-input-length-for-the-point.patch libgcrypt20-1.8.7/debian/patches/30_08-ecc-Check-the-input-length-for-the-point.patch
--- libgcrypt20-1.8.7/debian/patches/30_08-ecc-Check-the-input-length-for-the-point.patch	1970-01-01 01:00:00.0 +0100
+++ libgcrypt20-1.8.7/debian/patches/30_08-ecc-Check-the-input-length-for-the-point.patch	2021-05-02 13:52:32.0 +0200
@@ -0,0 +1,80 @@
+From 3f48e3ea37adf84aae7335b8367012d70bb3f132 Mon Sep 17 00:00:00 2001
+From: NIIBE Yutaka 
+Date: Tue, 27 Apr 2021 17:24:16 +0900
+Subject: [PATCH 8/8] ecc: Check the input length for the point.
+
+* cipher/ecc-misc.c (_gcry_ecc_mont_decodepoint): Check the length
+of valid point representation.
+
+--
+
+Backport the commit of master:
+
+	060c378c050e7ec6206358c681a313d6e1967dcf
+
+In the use case of GnuPG, ECDH decryption for anonymous recipient may
+try to decrypt with different curves.  When the input data of
+ephemeral key does not match one of the private key, it should return
+GPG_ERR_INV_OBJ.
+
+Signed-off-by: NIIBE Yutaka 
+---
+ cipher/ecc-misc.c | 18 ++
+ 1 file changed, 14 insertions(+), 4 deletions(-)
+
+diff --git a/cipher/ecc-misc.c b/cipher/ecc-misc.c
+index 34dd6804..b89dcfa6 100644
+--- a/cipher/ecc-misc.c
 b/cipher/ecc-misc.c
+@@ -294,6 +294,7 @@ _gcry_ecc_mont_decodepoint (gcry_mpi_t pk, mpi_ec_t ctx, mpi_point_t result)
+ {
+   unsigned char *rawmpi;
+   un

Bug#988508: buster-pu: package gnutls28/3.6.7-4+deb10u7

[Andreas Metzler]

b10u6),-] {+3.6.7-4+deb10u7),+} libgnutls30 (= [-3.6.7-4+deb10u6),-] {+3.6.7-4+deb10u7),+} libgnutlsxx28 (= [-3.6.7-4+deb10u6),-] {+3.6.7-4+deb10u7),+} libidn2-dev, libp11-kit-dev (>= 0.23.10), libtasn1-6-dev, nettle-dev (>= 3.4.1~rc1)
Installed-Size: [-4316-] {+4317+}
Version: [-3.6.7-4+deb10u6-] {+3.6.7-4+deb10u7+}

Control files of package libgnutls30: lines which differ (wdiff format)
---
Installed-Size: [-2648-] {+2649+}
Version: [-3.6.7-4+deb10u6-] {+3.6.7-4+deb10u7+}

Control files of package libgnutls30-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-bce53405fc0dd81c8428113ca15f83b4a0ef10de-] {+bf7d77672d078e8b8e4ce26a48e80e3770237d37+}
Depends: libgnutls30 (= [-3.6.7-4+deb10u6)-] {+3.6.7-4+deb10u7)+}
Version: [-3.6.7-4+deb10u6-] {+3.6.7-4+deb10u7+}

Control files of package libgnutlsxx28: lines which differ (wdiff format)
-
Depends: libgnutls30 (= [-3.6.7-4+deb10u6),-] {+3.6.7-4+deb10u7),+} libc6 (>= 2.14), libgcc1 (>= 1:3.0), libstdc++6 (>= 5)
Version: [-3.6.7-4+deb10u6-] {+3.6.7-4+deb10u7+}

Control files of package libgnutlsxx28-dbgsym: lines which differ (wdiff format)

Depends: libgnutlsxx28 (= [-3.6.7-4+deb10u6)-] {+3.6.7-4+deb10u7)+}
Version: [-3.6.7-4+deb10u6-] {+3.6.7-4+deb10u7+}



diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog	2021-01-02 18:10:33.0 +0100
+++ gnutls28-3.6.7/debian/changelog	2021-05-14 13:33:38.0 +0200
@@ -1,3 +1,25 @@
+gnutls28 (3.6.7-4+deb10u7) buster; urgency=medium
+
+  * 46_handshake-reject-no_renegotiation-alert-if-handshake.patch pulled from
+3.6.15: It was found by oss-fuzz that the server sending a
+"no_renegotiation" alert in an unexpected timing, followed by an invalid
+second handshake can cause a TLS 1.3 client to crash via a null-pointer
+dereference. The crash happens in the application's error handling path,
+where the gnutls_deinit function is called after detecting a handshake
+failure.
+GNUTLS-SA-2020-09-04 CVE-2020-24659 Closes: #969547
+  * Pull multiple fixes designated for 3.6.15 bugfix release:
++ 47_rel3.6.16_01-gnutls_buffer_append_data-remove-duplicated-code.patch
++ 47_rel3.6.16_02-_gnutls_buffer_resize-add-option-to-use-allocation-s.patch
++ 47_rel3.6.16_03-key_share-avoid-use-after-free-around-realloc.patch
+  (CVE-2021-20231) and
+  47_rel3.6.16_04-pre_shared_key-avoid-use-after-free-around-realloc.patch
+  (CVE-2021-20232), both together GNUTLS-SA-2021-03-10.
++ 47_rel3.6.16_05-_gnutls_buffer_resize-account-for-unused-area-if-AGG.patch
++ 47_rel3.6.16_06-str-suppress-Wunused-function-if-AGGRESSIVE_REALLOC-.patch
+
+ -- Andreas Metzler   Fri, 14 May 2021 13:33:38 +0200
+
 gnutls28 (3.6.7-4+deb10u6) buster; urgency=medium
 
   * 45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch
diff -Nru gnutls28-3.6.7/debian/patches/46_handshake-reject-no_renegotiation-alert-if-handshake.patch gnutls28-3.6.7/debian/patches/46_handshake-reject-no_renegotiation-alert-if-handshake.patch
--- gnutls28-3.6.7/debian/patches/46_handshake-reject-no_renegotiation-alert-if-handshake.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.6.7/debian/patches/46_handshake-reject-no_renegotiation-alert-if-handshake.patch	2021-05-11 18:08:00.0 +0200
@@ -0,0 +1,112 @@
+From 521e6492b9bbc8ec1519924526942cf2fc719497 Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Sat, 22 Aug 2020 17:19:39 +0200
+Subject: [PATCH] handshake: reject no_renegotiation alert if handshake is
+ incomplete
+
+If the initial handshake is incomplete and the server sends a
+no_renegotiation alert, the client should treat it as a fatal error
+even if its level is warning.  Otherwise the same handshake
+state (e.g., DHE parameters) are reused in the next gnutls_handshake
+call, if it is called in the loop idiom:
+
+  do {
+  ret = gnutls_handshake(session);
+  } while (ret < 0 && gnutls_error_is_fatal(ret) == 0);
+
+Signed-off-by: Daiki Ueno 
+---
+ ...a04b3d3f7dcc0ab4571cf0df3b67ab7e1005e9e7a8 | Bin 0 -> 671 bytes
+ ...1da801fb3c6d1f7f846f227721e221adea08aa319c | Bin 0 -> 729 bytes
+ lib/gnutls_int.h  |   1 +
+ lib/handshake.c   |  48 +-
+ 4 files changed, 36 insertions(+), 13 deletions(-)
+ create mode 100644 fuzz/gnutls_client_fuzzer.in/00ea40761ce11e769f1817a04b3d3f7dcc0ab4571cf0df3b67ab7e1005e9e7a8
+ create mode 100644 fuzz/gnutls_psk_client_fuzzer.in/b16434290b77e13d7a983d1da801fb3c6d1f7f846f227721e221adea08aa319c
+
+--- a/lib/gnutls_int.h
 b/lib/gnutls_int.h
+@@ -1366,6 +1366,8 @@ typedef struct {
+ #d

Bug#987924: unblock: exim4/4.94-19

[Andreas Metzler]

... forgot the debdiff.

cu Andreas
diff -Nru exim4-4.94/debian/changelog exim4-4.94/debian/changelog
--- exim4-4.94/debian/changelog 2021-03-18 13:54:47.0 +0100
+++ exim4-4.94/debian/changelog 2021-04-26 18:35:43.0 +0200
@@ -1,3 +1,42 @@
+exim4 (4.94-19) unstable; urgency=medium
+
+  * Further updates from heiko/exim-4.94+fixes+taintwarn:
++ 75_24-Silence-the-compiler.patch
++ 75_26-Disable-taintchecks-for-mkdir-this-isn-t-part-of-4.9.patch
+  * Upload to unstable.
+
+ -- Andreas Metzler   Mon, 26 Apr 2021 18:35:43 +0200
+
+exim4 (4.94-18) experimental; urgency=medium
+
+  * Pull patches to temporarily add an option to turn taint errors into
+warnings. (See #987133)
++ 75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
++ 75_02-search.patch
++ 75_03-dbstuff.patch
++ 75_04-acl.patch
++ 75_05-parse.patch
++ 75_06-rda.patch
++ 75_07-appendfile.patch
++ 75_08-autoreply.patch
++ 75_09-pipe.patch
++ 75_10-deliver.patch
++ 75_11-directory.patch
++ 75_12-expand.patch
++ 75_13-lf_sqlperform.patch
++ 75_14-rf_get_transport.patch
++ 75_15-deliver.patch
++ 75_16-smtp_out.patch
++ 75_17-smtp.patch
++ 75_18-update-doc.patch
++ 75_20-Set-mainlog_name-and-rejectlog_name-unconditionally.patch
++ 75_21-tidy-log.c.patch
++ 75_22-Silence-compiler.patch
++ 75_23-Do-not-close-the-main-_log-if-we-do-not-see-a-chance.patch
+  * Update NEWS.Debian to describe the feature.
+
+ -- Andreas Metzler   Sun, 25 Apr 2021 07:42:26 +0200
+
 exim4 (4.94-17) unstable; urgency=medium
 
   * Let exim4-config Recommend ca-certificates, needed for certificate
diff -Nru exim4-4.94/debian/NEWS exim4-4.94/debian/NEWS
--- exim4-4.94/debian/NEWS  2021-03-18 13:53:44.0 +0100
+++ exim4-4.94/debian/NEWS  2021-04-25 08:08:34.0 +0200
@@ -1,19 +1,10 @@
-exim4 (4.94-16) unstable; urgency=medium
+exim4 (4.94-18) experimental; urgency=medium
 
-  The configuration now enforces certificate verification against the
-  system trust store on encrypted connections using the
-  remote_smtp_smarthost transport (smarthost and satellite setups).
-  Delivery will therefore fail if the host certificates are not verifyable
-  and non TLS delivery is not available (e.g. because AUTH PLAIN is used).
-
- -- Andreas Metzler   Wed, 17 Mar 2021 13:50:44 +0100
-
-exim4 (4.94~RC0-2) experimental; urgency=low
-
-  Please consider this a *major* exim upgrade. It introduces the concept of
-  tainted data read from untrusted sources, like e.g. message sender or
-  recipient. This tainted data (e.g. $local_part or $domain) cannot be used
-  among other things as a file or directory name or command name.
+  Please consider exim 4.93/4.94 a *major* exim upgrade. It introduces the
+  concept of tainted data read from untrusted sources, like e.g. message
+  sender or recipient. This tainted data (e.g. $local_part or $domain)
+  cannot be used among other things as a file or directory name or command
+  name.
 
   This WILL BREAK configurations which are not updated accordingly.
   Old Debian exim configuration files also will not work unmodified, the new
@@ -32,7 +23,27 @@
   lookup in further processing instead of the original (remote provided)
   value.
 
- -- Andreas Metzler   Sun, 10 May 2020 10:27:04 +0200
+  To ease upgrading there is a new main configuration option to temporarily
+  downgrade taint errors to warnings, letting the old configuration work with
+  the newer exim. To make use of this feature add
+  .ifdef _OPT_MAIN_ALLOW_INSECURE_TAINTED_DATA
+   allow_insecure_tainted_data = yes
+  .endif
+  to the exim configuration (e.g. to /etc/exim4/exim4.conf.localmacros)
+  *before* upgrading to exim 4.93/4.94 and check the logfile for taint
+  warnings. This is a temporary workaround which will stop working in 4.95.
+
+ -- Andreas Metzler   Sun, 25 Apr 2021 07:42:26 +0200
+
+exim4 (4.94-16) unstable; urgency=medium
+
+  The configuration now enforces certificate verification against the
+  system trust store on encrypted connections using the
+  remote_smtp_smarthost transport (smarthost and satellite setups).
+  Delivery will therefore fail if the host certificates are not verifyable
+  and non TLS delivery is not available (e.g. because AUTH PLAIN is used).
+
+ -- Andreas Metzler   Wed, 17 Mar 2021 13:50:44 +0100
 
 exim4 (4.87-3) unstable; urgency=medium
 
diff -Nru 
exim4-4.94/debian/patches/75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
 
exim4-4.94/debian/patches/75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
--- 
exim4-4.94/debian/patches/75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
  1970-01-01 01:00:00.0 +0100
+++ 
exim4-4.94/debian/patches/75_01-Introduce-main-config-option-allow_insecure_tainted_.patch
  2021-04-25 07:27:42.0 +0200
@@ -0,0 +1,230 @@
+From ec06d64532e4952fc36429f73e0222d26997ef7c Mon Sep 17 00:00:00 2001
+From: "

Bug#987924: unblock: exim4/4.94-19

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: ex...@packages.debian.org

Hello,

Please consider exim4/4.94-19 for bullseye. Due to the newly introduced
tainting mechanism exim upgrades from buster to bullseye currently
require a lockstep upgrade of configuration file and exim binary. The new
binary will not run with the old configuration and vice versa. -19
brings a patch that adds an option ("allow_insecure_tainted_data") to let
the new daemon work with the old configuration at the price of spamming the
logfile.

-19 adds the option but has it disabled by default. I thought it fit
better to _require_ handholding /now/ at dist-upgrade time. I am open to
strong opinions to switch it on by default.

I would appreciate a timely feedback on this review request. There is a
security release scheduled for May 4th
https://lists.exim.org/lurker/message/20210421.123632.08bb711a.en.html
and I would like to be able prepare an upload and have it propagate to
testing ASAP (on the same day!) and not have to discuss
"allow_insecure_tainted_data", then. If you do not think that it is
possible/wise to accept allow_insecure_tainted_data that quickly I will
base the upload for the security release on -17 and will temporarily revert
17-->19.

unblock exim4/4.94-19

TIA, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#985984: unblock: exim4/4.94-17

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock
X-Debbugs-Cc: ex...@packages.debian.org

Hello release team,

Please unblock exim4.

The main point of this upload is to fix the issues reported by Jö Fahlke
which apply to bullseye. This was fixed for buster in a more conservative
way (document instead of improved behavior) in 4.92-8+deb10u5.

* README.Debian: Fix typo "tls_verify_certificate" instead of
  "tls_verify_certificates".
* General doc improvements in this area. (Thanks, Jö Fahlke) Closes: #985244
* Enforce certificate verification against the system trust store in the
  remote SMTP transport by default by setting
  REMOTE_SMTP_SMARTHOST_TLS_VERIFY_HOSTS = *. Closes: #985344
* Let exim4-config Recommend ca-certificates, needed for certificate
  verification.

The second important change is
* Intensify upgrade warning in NEWS file.

I have also synced with upstream's bugfix-only GIT branch.

unblock exim4/4.94-17

TIA, cu Andreas
diff -Nru exim4-4.94/debian/changelog exim4-4.94/debian/changelog
--- exim4-4.94/debian/changelog 2021-02-07 08:13:29.0 +0100
+++ exim4-4.94/debian/changelog 2021-03-18 13:54:47.0 +0100
@@ -1,3 +1,30 @@
+exim4 (4.94-17) unstable; urgency=medium
+
+  * Let exim4-config Recommend ca-certificates, needed for certificate
+verification.
+
+ -- Andreas Metzler   Thu, 18 Mar 2021 13:54:47 +0100
+
+exim4 (4.94-16) unstable; urgency=medium
+
+  * README.Debian: Fix typo "tls_verify_certificate" instead of
+"tls_verify_certificates".
+  * General doc improvements in this area. (Thanks, Jö Fahlke) Closes: #985244
+  * Intensify upgrade warning in NEWS file.
+  * Enforce certificate verification against the system trust store in the
+remote SMTP transport by default by setting
+REMOTE_SMTP_SMARTHOST_TLS_VERIFY_HOSTS = *. Closes: #985344
+  * Update from exim-4.94+fixes:
++ 74_56-Fix-FreeBSD-13-build.patch
++ 74_57-Fix-weight-calculation-for-spamd_address.-Bug-2694.patch
++ 74_58-Fix-weight-calculation-for-socks_proxy.-Bug-2694.patch
++ 74_59-Fix-build-for-platforms-not-having-ulong.patch
++ 74_60-Fix-list-expansion-for-various-domainlists-having-in.patch
++ 74_61-Bulid-fix-DISABLE_PIPE_CONNECT-build.-Bug-2703.patch
++ 74_62-Docs-fix-description-of-hosts_try_dane.-Bug-2704.patch
+
+ -- Andreas Metzler   Wed, 17 Mar 2021 13:50:44 +0100
+
 exim4 (4.94-15) unstable; urgency=medium
 
   * Update from exim-4.94+fixes:
diff -Nru exim4-4.94/debian/control exim4-4.94/debian/control
--- exim4-4.94/debian/control   2021-01-30 18:21:15.0 +0100
+++ exim4-4.94/debian/control   2021-03-18 13:54:47.0 +0100
@@ -109,6 +109,7 @@
  exim4-config-2,
  ${MTA-Conflicts}
 Depends: adduser, ${misc:Depends}, ${shlibs:Depends}
+Recommends: ca-certificates
 Description: configuration for the Exim MTA (v4)
  Exim (v4) is a mail transport agent. exim4-config provides the configuration
  for the exim4 daemon packages. The configuration framework has been split
diff -Nru 
exim4-4.94/debian/debconf/conf.d/transport/10_exim4-config_transport-macros 
exim4-4.94/debian/debconf/conf.d/transport/10_exim4-config_transport-macros
--- exim4-4.94/debian/debconf/conf.d/transport/10_exim4-config_transport-macros 
2020-02-29 15:37:28.0 +0100
+++ exim4-4.94/debian/debconf/conf.d/transport/10_exim4-config_transport-macros 
2021-03-18 13:53:44.0 +0100
@@ -14,3 +14,7 @@
 REMOTE_SMTP_HELO_DATA=${lookup dnsdb 
{ptr=$sending_ip_address}{$value}{$primary_hostname}}
 .endif
 .endif
+
+.ifndef REMOTE_SMTP_SMARTHOST_TLS_VERIFY_HOSTS
+  REMOTE_SMTP_SMARTHOST_TLS_VERIFY_HOSTS = *
+.endif
diff -Nru exim4-4.94/debian/NEWS exim4-4.94/debian/NEWS
--- exim4-4.94/debian/NEWS  2020-11-03 18:11:40.0 +0100
+++ exim4-4.94/debian/NEWS  2021-03-18 13:53:44.0 +0100
@@ -1,11 +1,36 @@
+exim4 (4.94-16) unstable; urgency=medium
+
+  The configuration now enforces certificate verification against the
+  system trust store on encrypted connections using the
+  remote_smtp_smarthost transport (smarthost and satellite setups).
+  Delivery will therefore fail if the host certificates are not verifyable
+  and non TLS delivery is not available (e.g. because AUTH PLAIN is used).
+
+ -- Andreas Metzler   Wed, 17 Mar 2021 13:50:44 +0100
+
 exim4 (4.94~RC0-2) experimental; urgency=low
 
-  Some Transports now refuse to use tainted data in constructing their
-  delivery location; this WILL BREAK configurations which are not updated
-  accordingly.  In particular: any Transport use of $local_user which has
-  been relying upon check_local_user far away in the Router to make it
-  safe, should be updated to replace $local_user with
-  $local_part_data.
+  Please consider this a *major* exim upgrade. It introduces the concept of
+  tainted data read from untrusted sources, like e.g. message sender or
+  recipient. This tainted data (e.g. $local_part or $doma

Bug#985466: buster-pu: package libpano13/2.9.19+dfsg-3+deb10u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: libpan...@packages.debian.org

Hello,

I would like to fix 985249 for buster. It is a straightforward format
string issue, as documented in the respective report. The issue is fixed
in unstable (2.9.20~rc3+dfsg-1) but not yet in testing.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru libpano13-2.9.19+dfsg/debian/changelog 
libpano13-2.9.19+dfsg/debian/changelog
--- libpano13-2.9.19+dfsg/debian/changelog  2017-09-10 14:39:18.0 
+0200
+++ libpano13-2.9.19+dfsg/debian/changelog  2021-03-18 14:12:08.0 
+0100
@@ -1,3 +1,12 @@
+libpano13 (2.9.19+dfsg-3+deb10u1) buster; urgency=medium
+
+  * 850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff
+cherry-picked from 2.9.20 rc3: Fixes format string bug, pasing along
+format strings in user specified output filename to printf.
+Closes: #985249
+
+ -- Andreas Metzler   Thu, 18 Mar 2021 14:12:08 +0100
+
 libpano13 (2.9.19+dfsg-3) unstable; urgency=medium
 
   * Move Vcs-* from git/http to https.
diff -Nru 
libpano13-2.9.19+dfsg/debian/patches/850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff
 
libpano13-2.9.19+dfsg/debian/patches/850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff
--- 
libpano13-2.9.19+dfsg/debian/patches/850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff
 1970-01-01 01:00:00.0 +0100
+++ 
libpano13-2.9.19+dfsg/debian/patches/850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff
 2021-03-18 14:12:08.0 +0100
@@ -0,0 +1,40 @@
+# HG changeset patch
+# User tmodes
+# Date 1615911819 -3600
+#  Tue Mar 16 17:23:39 2021 +0100
+# Node ID f02459498cb44c0087900616a7e61563d614c05f
+# Parent  2e9ee0a5e32f2ca6e1a5b3f9c2d5c393a41903c3
+Prevent string vulnerability by refusing prefix strings with percentage sign
+
+diff -r 2e9ee0a5e32f -r f02459498cb4 file.c
+--- a/file.c   Sun Dec 13 15:37:56 2020 +0100
 b/file.c   Tue Mar 16 17:23:39 2021 +0100
+@@ -2910,6 +2910,16 @@
+ }
+ strcat(outputPrefix, DEFAULT_PREFIX_NUMBER_FORMAT);
+ }
++else {
++// TODO: sanitize outputPrefix, only a single format specifier %??d 
or %??i
++// is allowed, all other should be escaped
++// until this is implemented refuse to process further if prefix 
string
++// contains a percentage sign to prevent string vulnerability in 
++// sprintf(outputFilename, outputPrefix ...) below
++PrintError("Output prefix must not contain a percentage sign");
++return 0;
++}
++
+ 
+ for (i =0; i< filesCount ; i++) {
+ sprintf( outputFilename, outputPrefix, i );
+diff -r 2e9ee0a5e32f -r f02459498cb4 tools/PTcrop.c
+--- a/tools/PTcrop.c   Sun Dec 13 15:37:56 2020 +0100
 b/tools/PTcrop.c   Tue Mar 16 17:23:39 2021 +0100
+@@ -36,7 +36,7 @@
+ 
+ #define PT_CROP_USAGE "PTuncrop [options] \n\n" \
+ "Options:\n"\
+-"\t-p \tPrefix for output files (defaults to " DEFAULT_PREFIX 
"%%4d)\n"   \
++"\t-p \tPrefix for output files (defaults to " DEFAULT_PREFIX 
")\n"   \
+ "\t-f\t\tForce processing: Overwrite output files if they exists (use 
with care)\n"  \
+ "\t-x\t\tDelete source files (use with care)\n"\
+ "\t-q\t\tQuiet run\n"\
diff -Nru libpano13-2.9.19+dfsg/debian/patches/series 
libpano13-2.9.19+dfsg/debian/patches/series
--- libpano13-2.9.19+dfsg/debian/patches/series 1970-01-01 01:00:00.0 
+0100
+++ libpano13-2.9.19+dfsg/debian/patches/series 2021-03-18 14:12:08.0 
+0100
@@ -0,0 +1 @@
+850_f02459498cb4_Prevent_string_vulnerability_by_refusing.diff


signature.asc
Description: PGP signature

Bug#985450: buster-pu: package exim4/4.92-8+deb10u5

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: ex...@packages.debian.org, Jorrit Fahlke 

Hello,

[ Reason ]
I would like to fix two issues in buster:
#1 Fix use of concurrent TLS connections under GnuTLS.  When a callout was
   done during a receiving connection, and both used TLS, global info was
   used rather than per-connection info for tracking the state of data
   queued for transmission.  This could result in a connection hang.

#2 Fix issues related to certificate checking:
a) Cherry-pick a bugfix to get proper hostname checking with CNAMES.
Without this patch when connecting to a CNAME the server provided cert
is checked against the A record instead of the original cname. #985243

b) Document limitation/extent of server certificate checking that is
done by default and how to change it.  #985244 and #985344

2a and 2b are documented in the respective bug-reports, 2a actually
might warant priority serious. #1 has popped up repeatedly on the
exim-user mailing list, I would categorize it as important.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.92/debian/changelog exim4-4.92/debian/changelog
--- exim4-4.92/debian/changelog 2020-05-13 18:01:31.0 +0200
+++ exim4-4.92/debian/changelog 2021-03-18 09:10:15.0 +0100
@@ -1,3 +1,23 @@
+exim4 (4.92-8+deb10u5) buster; urgency=medium
+
+  * Fix use of concurrent TLS connections under GnuTLS:
+80_01-GnuTLS-fix-hanging-callout-connections.patch
+80_02-GnuTLS-tls_write-wait-after-uncorking-the-session.patch
+80_03-GnuTLS-Do-not-care-about-corked-data-when-uncorking.patch
+(Thanks, Heiko Schlittermann for the backport)
+  * Pull 82_TLS-use-RFC-6125-rules-for-certifucate-name-checks-w.patch from
+upstream git (already included in 4.94), on TLS connections to a CNAME
+verify the certificate against the original CNAME instead of against
+the A record. Closes: #985243
+  * In README.Debian explicitly document the limitation/extent of server
+certificate checking (authenticity not enforced) in the default
+configuration (Thanks, Jö Fahlke). This Closes: #985244 (improved
+documentation and Closes: #985344 (Yes, without required cert
+checking MitM attacks are possible, but for a stable update documenting
+this is the best compromise.)
+
+ -- Andreas Metzler   Thu, 18 Mar 2021 09:10:15 +0100
+
 exim4 (4.92-8+deb10u4) buster-security; urgency=high
 
   * Fix authentication bypass in SPA authenticator due to out-of-bound buffer
diff -Nru 
exim4-4.92/debian/patches/80_01-GnuTLS-fix-hanging-callout-connections.patch 
exim4-4.92/debian/patches/80_01-GnuTLS-fix-hanging-callout-connections.patch
--- 
exim4-4.92/debian/patches/80_01-GnuTLS-fix-hanging-callout-connections.patch
1970-01-01 01:00:00.0 +0100
+++ 
exim4-4.92/debian/patches/80_01-GnuTLS-fix-hanging-callout-connections.patch
2021-03-18 08:51:35.0 +0100
@@ -0,0 +1,83 @@
+From 97c5e07c220b55d1c506a1798c9ce3ae3105adea Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Thu, 13 Feb 2020 16:45:38 +
+Subject: [PATCH 4/6] GnuTLS: fix hanging callout connections
+
+Broken-by: 925ac8e4f1
+(cherry picked from commit bd95ffc2ba87fbd3c752df17bc8fd9c01586d45a)
+---
+ doc/ChangeLog | 81 ---
+ src/tls-gnu.c | 24 +++--
+ 2 files changed, 20 insertions(+), 85 deletions(-)
+
+--- a/doc/ChangeLog
 b/doc/ChangeLog
+@@ -67,6 +67,11 @@ JH/41 Bug 2571: Fix SPA authenticator.
+   being used.  A malicious client could thus cause an out-of-bounds read 
and
+   possibly gain authentication.  Fix by adding the check.
+ 
++JH/25 Fix use of concurrent TLS connections under GnuTLS.  When a callout was
++  done during a receiving connection, and both used TLS, global info was
++  used rather than per-connection info for tracking the state of data
++  queued for transmission.  This could result in a connection hang.
++
+ 
+ Exim version 4.92
+ -
+--- a/src/tls-gnu.c
 b/src/tls-gnu.c
+@@ -124,10 +124,17 @@ typedef struct exim_gnutls_state {
+   enum peer_verify_requirement verify_requirement;
+   int fd_in;
+   int fd_out;
+-  BOOLpeer_cert_verified;
+-  BOOLpeer_dane_verified;
+-  BOOLtrigger_sni_changes;
+-  BOOLhave_set_peerdn;
++
++  BOOLpeer_cert_verified:1;
++  BOOLpeer_dane_verified:1;
++  BOOLtrigger_sni_changes:1;
++  BOOLhave_set_peerdn:1;
++  BOOLxfer_eof:1; /*XXX never gets set! */
++  BOOLxfer_error:1;
++#ifdef SUPPORT_CORK
++  BOOLcorked:1

Bug#981581: nmu: sa-exim_4.2.1-19

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: binnmu

nmu sa-exim_4.2.1-19 . ANY . unstable . -m "Rebuild against current exim 
localscan ABI. (See #981398)"

That is necessary to let sa-exim work again in sid and bulleye. The
wrong Provides of exim4-daemon-light in bullseye (#981399) allow
installation of sa-exim, but it does not work. Current exim4 fixes this.

Details in https://bugs.debian.org/981398

TIA, cu Andreas

Bug#979074: buster-pu: package gnutls28/3.6.7-4+deb10u6

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu

[ Reason ]
The gnutls28 test tests/testpkcs11.sh uses a test certificate that
expired in December 2020, which now causes a testsuite error and FTBFS.
If this is not approved the patch will need to be included in case of
another DSA for GnuTLS or a stable update. I would rather fix this now
to make debian-security's life easier.

[ Checklist ]
  [X] *all* changes are documented in the d/changelog
  [x] I reviewed all changes and I approve them
  [x] attach debdiff against the package in (old)stable
  [x] the issue is verified as fixed in unstable

[ Changes ]
The patch uses datefudge to avoid the timebomb. It is cherrypicked and
adapted (older helper function) from upstream master.

TIA, cu Andreas

diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog	2020-06-07 07:45:55.0 +0200
+++ gnutls28-3.6.7/debian/changelog	2021-01-02 14:15:36.0 +0100
@@ -1,3 +1,11 @@
+gnutls28 (3.6.7-4+deb10u6) UNRELEASED; urgency=medium
+
+  * 45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch
+Fix test suite error caused by expired certificate.
+Closes: #977552
+
+ -- Andreas Metzler   Sat, 02 Jan 2021 14:15:36 +0100
+
 gnutls28 (3.6.7-4+deb10u5) buster; urgency=medium
 
   * 42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch
diff -Nru gnutls28-3.6.7/debian/patches/45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch gnutls28-3.6.7/debian/patches/45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch
--- gnutls28-3.6.7/debian/patches/45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.6.7/debian/patches/45_4.7.0plus-01_testpkcs11-use-datefudge-to-trick-certificate-expiry.patch	2021-01-02 14:15:36.0 +0100
@@ -0,0 +1,73 @@
+From 2b0f6f3a2ff13153aaa70c764ba7a8b90aef794d Mon Sep 17 00:00:00 2001
+From: Daiki Ueno 
+Date: Mon, 28 Dec 2020 16:16:53 +0100
+Subject: [PATCH 3/6] testpkcs11: use datefudge to trick certificate expiry
+Origin: https://gitlab.com/gnutls/gnutls/-/commit/2b0f6f3a2ff13153aaa70c764ba7a8b90aef794d
+Bug: https://gitlab.com/gnutls/gnutls/-/issues/1135
+Bug-Debian: https://bugs.debian.org/977552
+
+The certificates stored in tests/testpkcs11-certs expired on
+2020-12-13.  To avoid verification failure due to that, use datefudge
+to set custom date when calling gnutls-cli, gnutls-serv, and certtool.
+
+Based on the patch by Andreas Metzler:
+https://gitlab.com/gnutls/gnutls/-/issues/1135#note_469682121
+
+Signed-off-by: Daiki Ueno 
+---
+ tests/testpkcs11.sh | 12 +++-
+ 1 file changed, 11 insertions(+), 1 deletion(-)
+
+--- a/tests/testpkcs11.sh
 b/tests/testpkcs11.sh
+@@ -67,6 +67,8 @@ have_ed25519=0
+ P11TOOL="${VALGRIND} ${P11TOOL} --batch"
+ SERV="${SERV} -q"
+ 
++TESTDATE=2020-12-01
++
+ . ${srcdir}/scripts/common.sh
+ 
+ rm -f "${LOGFILE}"
+@@ -79,6 +81,8 @@ exit_error () {
+ 	exit 1
+ }
+ 
++skip_if_no_datefudge
++
+ # $1: token
+ # $2: PIN
+ # $3: filename
+@@ -510,6 +514,7 @@ write_certificate_test () {
+ 	pubkey="$5"
+ 
+ 	echo -n "* Generating client certificate... "
++	datefudge -s "$TESTDATE" \
+ 	"${CERTTOOL}" ${CERTTOOL_PARAM} ${ADDITIONAL_PARAM}  --generate-certificate --load-ca-privkey "${cakey}"  --load-ca-certificate "${cacert}"  \
+ 	--template ${srcdir}/testpkcs11-certs/client-tmpl --load-privkey "${token};object=gnutls-client;object-type=private" \
+ 	--load-pubkey "$pubkey" --outfile tmp-client.crt >>"${LOGFILE}" 2>&1
+@@ -887,6 +892,7 @@ use_certificate_test () {
+ 	echo -n "* Using PKCS #11 with gnutls-cli (${txt})... "
+ 	# start server
+ 	eval "${GETPORT}"
++	SERV="datefudge -s $TESTDATE $SERV" \
+ 	launch_pkcs11_server $$ "${ADDITIONAL_PARAM}" --echo --priority NORMAL --x509certfile="${certfile}" \
+ 		--x509keyfile="$keyfile" --x509cafile="${cafile}" \
+ 		--verify-client-cert --require-client-cert >>"${LOGFILE}" 2>&1
+@@ -895,13 +901,16 @@ use_certificate_test () {
+ 	wait_server ${PID}
+ 
+ 	# connect to server using SC
++	datefudge -s "$TESTDATE" \
+ 	${VALGRIND} "${CLI}" ${ADDITIONAL_PARAM} -p "${PORT}" localhost --priority NORMAL --x509cafile="${cafile}" >"${LOGFILE}" 2>&1 && \
+ 		fail ${PID} "Connection should have failed!"
+ 
++	datefudge -s "$TESTDATE" \
+ 	${VALGRIND} "${CLI}" ${ADDITIONAL_PARAM} -p "${PORT}" localhost --priority NORMAL --x509certfile="${certfile}" \
+ 	--x509keyfile="$keyfile" --x509cafile="${cafile}" >"${LOGFILE}" 2

Bug#963703: stretch-pu: package gnutls28/3.5.8-5+deb9u5

[Andreas Metzler]

On 2020-07-02 "Adam D. Barratt"  wrote:
> Control: tags -1 + confirmed

> On Thu, 2020-06-25 at 17:37 +0200, Andreas Metzler wrote:
>> I would like to make a last bugfix upload to stretch:
>> * Pull fixes for CVE-2019-3836 / [GNUTLS-SA-2019-03-27, #694].
[...]
> Please go ahead.

Thanks, done.

Bug#963703: stretch-pu: package gnutls28/3.5.8-5+deb9u5

[Andreas Metzler]

On 2020-06-25 Salvatore Bonaccorso  wrote:
> On Thu, Jun 25, 2020 at 05:41:42PM +0200, Andreas Metzler wrote:
> > On 2020-06-25 Andreas Metzler  wrote:
> > [...]
> > > * Pull fixes for CVE-2019-3836 / [GNUTLS-SA-2019-03-27, #694].
[...]
> Only noticed too late, but the above CVE reference should be
> CVE-2019-3829 in the "Pull fixes for ..." changelog entry.

Thanks, you are right. (The other CVE is also GNUTLS-SA-2019-03-27,
but the other upstream bug number #704.)

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#963703: stretch-pu: package gnutls28/3.5.8-5+deb9u5

[Andreas Metzler]

On 2020-06-25 Andreas Metzler  wrote:
[...]
> * Pull fixes for CVE-2019-3836 / [GNUTLS-SA-2019-03-27, #694].
>   + 40_casts_related_to_fix_CVE-2019-3829.patch
>   + 40_rel3.6.7_01-Automatically-NULLify-after-gnutls_free.patch
>   + 40_rel3.6.7_01-fuzz-added-fuzzer-for-certificate-verification.patch
>   + 41_use_datefudge_to_trigger_CVE-2019-3829_testcase.diff
> * More important fixes:
>   + 43_rel3.6.14_10-session_pack-fix-leak-in-error-path.patch
> [One-line-fix for memleak]

>   + 44_rel3.6.14_10-Update-session_ticket.c-to-add-support-for-zero-leng.patch
> Handle zero length session tickets, fixing connection errors on TLS1.2
> sessions to some big hosting providers. (See LP 1876286)
> [Fixes connections to e.g. verizon popserver.]
[...]

Here is the missing debdiff.

cu Andreas
[The following lists of changes regard files as different if they have
different names, permissions or owners.]

Files only in first set of .debs, found in package gnutls-bin-dbgsym

-rw-r--r--  root/root   /usr/lib/debug/.build-id/0e/df33e82a82671f7e361a8ffa83b02400337604.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/1d/b976be2d75d79dfd97e68dba3ee84babe5a3cc.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/64/414524cec63b3a8334146aa0c4dab71fae4080.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/6f/0012f94a9f80ef7e652dacc713347841f66907.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/98/eef0a29dcce526336be09fbbb0eccb3ece9f17.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/a5/c92e78a7d0a175b524703387c994518830abfa.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/ad/42bf08cf713e4a18ed1dd04dcc200a1cdafe94.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/c0/cf4951b3020f4fdf0b30c32934e922348e3660.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/f7/a745a4765a1efbfc31d0e21d0b5aca9aa2c5b1.debug

Files only in first set of .debs, found in package libgnutls-dane0-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/1c/399494f95f5e9ff28fcbd0243e96639fad69d3.debug

Files only in first set of .debs, found in package libgnutls-openssl27-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/51/a6d9549543590e69584a2dd9df4e919cd62918.debug

Files only in first set of .debs, found in package libgnutls30-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/1c/1bc93c559cfe2ebd1b5676fa4b355118edf38e.debug

Files only in first set of .debs, found in package libgnutlsxx28-dbgsym
---
-rw-r--r--  root/root   /usr/lib/debug/.build-id/f4/43a08baf0b78f1286c82e9d3e085c83734d37b.debug

New files in second set of .debs, found in package gnutls-bin-dbgsym

-rw-r--r--  root/root   /usr/lib/debug/.build-id/41/3e7554b4f2cfebbd3c79dbdc11815b1a8ce65b.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/4a/7b934e15748037c09f179e902c900edd8f645e.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/84/79ca3705d519462b64cafa740069f5257a1799.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/96/f89b1b2de8078f07e1dceb9a1c9570ce2fefe8.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/9c/a5be6ce0e2ad9359bcc4fa67713fb35451eb4f.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/d9/a524219c966f40c7f8862e5141f95e747ffa87.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/ea/4ce0554a6816fbe557433397e8a3f211063cb0.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/f0/bb902cc0ca2f32c8dbfc88908486ac7a52ca09.debug
-rw-r--r--  root/root   /usr/lib/debug/.build-id/f0/c69abf17ada32042b18f4b6b240c76965fe456.debug

New files in second set of .debs, found in package libgnutls-dane0-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/a2/c2822b8a615e4e750944c36cfa27e4c39d5448.debug

New files in second set of .debs, found in package libgnutls-openssl27-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/2a/a940233375f7c77955565715aa6404a4334c07.debug

New files in second set of .debs, found in package libgnutls30-dbgsym
-
-rw-r--r--  root/root   /usr/lib/debug/.build-id/b6/c9dec4e163583c6c1f2ea1b1ef75b1db2e6a0c.debug

New files in second set of .debs, found in package libgnutlsxx28-dbgsym
---
-rw-r--r--  root/root   /usr/lib/debug/.build-id/e2/00ca7e603a3c2ea4f81f9542ab13919b24b73e.debug


Control files of package gnutls-b

Bug#963703: stretch-pu: package gnutls28/3.5.8-5+deb9u5

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

Hello,

I would like to make a last bugfix upload to stretch:
* Pull fixes for CVE-2019-3836 / [GNUTLS-SA-2019-03-27, #694].
  + 40_casts_related_to_fix_CVE-2019-3829.patch
  + 40_rel3.6.7_01-Automatically-NULLify-after-gnutls_free.patch
  + 40_rel3.6.7_01-fuzz-added-fuzzer-for-certificate-verification.patch
  + 41_use_datefudge_to_trigger_CVE-2019-3829_testcase.diff
* More important fixes:
  + 43_rel3.6.14_10-session_pack-fix-leak-in-error-path.patch
[One-line-fix for memleak]

  + 44_rel3.6.14_10-Update-session_ticket.c-to-add-support-for-zero-leng.patch
Handle zero length session tickets, fixing connection errors on TLS1.2
sessions to some big hosting providers. (See LP 1876286)
[Fixes connections to e.g. verizon popserver.]

TIA, cu Andreas

Bug#960836: buster-pu: package gnutls28/3.6.7-4+deb10u4

[Andreas Metzler]

Control: tags -1 - moreinfo
Control: retitle -1 buster-pu: package gnutls28/3.6.7-4+deb10u5

On 2020-05-26 Andreas Metzler  wrote:
> Control: tags 960836 + moreinfo

> Please hold on approving this. I will probably need to add a fix for
> https://gitlab.com/gnutls/gnutls/-/issues/997

Hello,

find attached a new version rebased on the latests DSA and featuring these
additional fixes:

* 44_rel3.6.14_10-Update-session_ticket.c-to-add-support-for-zero-leng.patch
  from GnuTLS 3.6.14: Handle zero length session tickets, fixing connection
  errors on TLS1.2 sessions to some big hosting providers. (See LP 1876286)
* 44_rel3.6.14_15-_gnutls_pkcs11_verify_crt_status-check-validity-agai.patch
  44_rel3.6.14_16-x509-trigger-fallback-verification-path-when-cert-is.patch
  44_rel3.6.14_17-tests-add-test-case-for-certificate-chain-supersedin.patch
  backported from GnuTLS 3.6.14: Fix verification error with alternate
  chains. Closes: #961889

TIA, cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog	2020-06-05 19:32:17.0 +0200
+++ gnutls28-3.6.7/debian/changelog	2020-06-07 07:45:55.0 +0200
@@ -1,3 +1,24 @@
+gnutls28 (3.6.7-4+deb10u5) buster; urgency=medium
+
+  * 42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch
+from GNUTLS 3.6.11: Fix TL1.2 resumption errors. Closes: #956649
+  * 47_rel3.6.13_10-session_pack-fix-leak-in-error-path.patch from GNUTLS
+3.6.14: One line fix for memory leak. Closes: #958704
+  * Rename
+44_rel3.6.14_01-stek-differentiate-initial-state-from-valid-time-win.patch
+(security upload) to 44_rel3.6.14_90_... to be able to pull earlier fixes
+from 3.6.14 and have correct patch filename order.
+  * 44_rel3.6.14_10-Update-session_ticket.c-to-add-support-for-zero-leng.patch
+from GnuTLS 3.6.14: Handle zero length session tickets, fixing connection
+errors on TLS1.2 sessions to some big hosting providers. (See LP 1876286)
+  * 44_rel3.6.14_15-_gnutls_pkcs11_verify_crt_status-check-validity-agai.patch
+44_rel3.6.14_16-x509-trigger-fallback-verification-path-when-cert-is.patch
+44_rel3.6.14_17-tests-add-test-case-for-certificate-chain-supersedin.patch
+backported from GnuTLS 3.6.14: Fix verification error with alternate
+chains. Closes: #961889
+
+ -- Andreas Metzler   Sun, 07 Jun 2020 07:45:55 +0200
+
 gnutls28 (3.6.7-4+deb10u4) buster-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch
--- gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch	2020-06-07 06:48:47.0 +0200
@@ -0,0 +1,610 @@
+From afa6e340c084542ef416afc96dd0329f5507 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos 
+Date: Tue, 8 Oct 2019 07:23:31 +0200
+Subject: [PATCH] session tickets: parse extension during session resumption on
+ client side
+
+It is possible for a server to send a new session ticket during
+TLS1.2 resumption. To be able to parse it as client we need to
+check the extension during resumption as well.
+
+Resolves: #841
+
+Signed-off-by: Nikos Mavrogiannopoulos 
+---
+ NEWS|  3 +++
+ lib/ext/alpn.c  |  3 ++-
+ lib/ext/client_cert_type.c  |  3 ++-
+ lib/ext/cookie.c|  3 ++-
+ lib/ext/dumbfw.c|  3 ++-
+ lib/ext/early_data.c|  3 ++-
+ lib/ext/ec_point_formats.c  |  3 ++-
+ lib/ext/etm.c   |  3 ++-
+ lib/ext/ext_master_secret.c |  3 ++-
+ lib/ext/heartbeat.c |  3 ++-
+ lib/ext/key_share.c |  3 ++-
+ lib/ext/max_record.c|  3 ++-
+ lib/ext/post_handshake.c|  3 ++-
+ lib/ext/pre_shared_key.c|  3 ++-
+ lib/ext/psk_ke_modes.c  |  3 ++-
+ lib/ext/record_size_limit.c |  3 ++-
+ lib/ext/safe_renegotiation.c|  3 ++-
+ lib/ext/server_cert_type.c  |  3 ++-
+ lib/ext/server_name.c   |  3 ++-
+ lib/ext/session_ticket.c|  7 ++-
+ lib/ext/signature.c |  3 ++-
+ lib/ext/srp.c   |  3 ++-
+ lib/ext/srtp.c  |  3 ++-
+ lib/ext/status_request.c|  3 ++-
+ lib/ext/supported_groups.c  |  3 ++-
+ lib/ext/supported_versions.c|  3 ++-
+ lib/hello_ext.c | 36 ++---
+ lib/hello_ext.h |  3 ++-
+ lib/includes/gnutls/gnutls.h.in |  4 ++--
+ tests/gnutls-cli-resume.sh 

Bug#960836: buster-pu: package gnutls28/3.6.7-4+deb10u4

[Andreas Metzler]

Control: tags 960836 + moreinfo

Please hold on approving this. I will probably need to add a fix for
https://gitlab.com/gnutls/gnutls/-/issues/997

cu Andreas

Bug#960836: buster-pu: package gnutls28/3.6.7-4+deb10u4

[Andreas Metzler]

Control: tags 960836 - moreinfo

On 2020-05-17 "Adam D. Barratt"  wrote:
> Control: tags -1 + moreinfo

> On Sun, 2020-05-17 at 14:23 +0200, Andreas Metzler wrote:
>> I would like to update gnutls to fix #95664 aka
>> https://gitlab.com/gnutls/gnutls/-/issues/841 fixing TLS1.2 client
>> side resumption errors.

> #956649. :-)

> I'm assuming this is fixed in at least unstable already, but the BTS
> metadata suggests otherwise (potentially not helped by the local
> "found" version).

> Please could you confirm, and fix either the metadata or unstable.

Hello Adam,

Yes, it is fixed in both buster and sid, I have corrected the bug
metadata accordingly.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#960836: buster-pu: package gnutls28/3.6.7-4+deb10u4

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu

Hello,

I would like to update gnutls to fix #95664 aka
https://gitlab.com/gnutls/gnutls/-/issues/841 fixing TLS1.2 client side
resumption errors. And while I am at it I would also pick a one-line
fix for a memory leak (Fix requested in #958704.)

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog	2020-04-03 21:31:50.0 +0200
+++ gnutls28-3.6.7/debian/changelog	2020-05-17 13:45:29.0 +0200
@@ -1,3 +1,12 @@
+gnutls28 (3.6.7-4+deb10u4) buster; urgency=medium
+
+  * 42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch
+from GNUTLS 3.6.11: Fix TL1.2 resumption errors. Closes: #956649
+  * 47_rel3.6.14_10-session_pack-fix-leak-in-error-path.patch from GNUTLS
+3.6.14: One line fix for memory leak. Closes: #958704
+
+ -- Andreas Metzler   Sun, 17 May 2020 13:45:29 +0200
+
 gnutls28 (3.6.7-4+deb10u3) buster-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch
--- gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.6.7/debian/patches/42_rel3.6.11_10-session-tickets-parse-extension-during-session-resum.patch	2020-05-17 10:08:09.0 +0200
@@ -0,0 +1,610 @@
+From afa6e340c084542ef416afc96dd0329f5507 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos 
+Date: Tue, 8 Oct 2019 07:23:31 +0200
+Subject: [PATCH] session tickets: parse extension during session resumption on
+ client side
+
+It is possible for a server to send a new session ticket during
+TLS1.2 resumption. To be able to parse it as client we need to
+check the extension during resumption as well.
+
+Resolves: #841
+
+Signed-off-by: Nikos Mavrogiannopoulos 
+---
+ NEWS|  3 +++
+ lib/ext/alpn.c  |  3 ++-
+ lib/ext/client_cert_type.c  |  3 ++-
+ lib/ext/cookie.c|  3 ++-
+ lib/ext/dumbfw.c|  3 ++-
+ lib/ext/early_data.c|  3 ++-
+ lib/ext/ec_point_formats.c  |  3 ++-
+ lib/ext/etm.c   |  3 ++-
+ lib/ext/ext_master_secret.c |  3 ++-
+ lib/ext/heartbeat.c |  3 ++-
+ lib/ext/key_share.c |  3 ++-
+ lib/ext/max_record.c|  3 ++-
+ lib/ext/post_handshake.c|  3 ++-
+ lib/ext/pre_shared_key.c|  3 ++-
+ lib/ext/psk_ke_modes.c  |  3 ++-
+ lib/ext/record_size_limit.c |  3 ++-
+ lib/ext/safe_renegotiation.c|  3 ++-
+ lib/ext/server_cert_type.c  |  3 ++-
+ lib/ext/server_name.c   |  3 ++-
+ lib/ext/session_ticket.c|  7 ++-
+ lib/ext/signature.c |  3 ++-
+ lib/ext/srp.c   |  3 ++-
+ lib/ext/srtp.c  |  3 ++-
+ lib/ext/status_request.c|  3 ++-
+ lib/ext/supported_groups.c  |  3 ++-
+ lib/ext/supported_versions.c|  3 ++-
+ lib/hello_ext.c | 36 ++---
+ lib/hello_ext.h |  3 ++-
+ lib/includes/gnutls/gnutls.h.in |  4 ++--
+ tests/gnutls-cli-resume.sh  | 17 
+ 30 files changed, 98 insertions(+), 44 deletions(-)
+
+ 
+diff --git a/lib/ext/alpn.c b/lib/ext/alpn.c
+index b9991f0a1..7cc799756 100644
+--- a/lib/ext/alpn.c
 b/lib/ext/alpn.c
+@@ -39,7 +39,8 @@ const hello_ext_entry_st ext_mod_alpn = {
+ 	.tls_id = 16,
+ 	.gid = GNUTLS_EXTENSION_ALPN,
+ 	/* this extension must be parsed even on resumption */
+-	.parse_type = GNUTLS_EXT_MANDATORY,
++	.client_parse_point = GNUTLS_EXT_MANDATORY,
++	.server_parse_point = GNUTLS_EXT_MANDATORY,
+ 	.validity = GNUTLS_EXT_FLAG_TLS | GNUTLS_EXT_FLAG_DTLS |
+ 		GNUTLS_EXT_FLAG_CLIENT_HELLO | GNUTLS_EXT_FLAG_EE |
+ 		GNUTLS_EXT_FLAG_TLS12_SERVER_HELLO,
+diff --git a/lib/ext/client_cert_type.c b/lib/ext/client_cert_type.c
+index b627b71f9..34f4dcfa4 100644
+--- a/lib/ext/client_cert_type.c
 b/lib/ext/client_cert_type.c
+@@ -48,7 +48,8 @@ const hello_ext_entry_st ext_mod_client_cert_type = {
+ 	.name = "Client Certificate Type",
+ 	.tls_id = 19,
+ 	.gid = GNUTLS_EXTENSION_CLIENT_CERT_TYPE,
+-	.parse_type = GNUTLS_EXT_TLS,
++	.client_parse_point = GNUTLS_EXT_TLS,
++	.server_parse_point = GNUTLS_EXT_TLS,
+ 	.validity = GNUTLS_EXT_FLAG_TLS |
+ 		GNUTLS_EXT_FLAG_DTLS |
+ 		GNUTLS_EXT_FLAG_CLIENT_HELLO |
+diff --git a/lib/ext/cookie.c b/lib/ext/cookie.c
+index 0feb2f0e5..b4608f3a9 100644
+--- a/lib/ext/cookie.c
 b/lib/ext/cookie.c
+@@ -41,7 +41,8 @@ const hello_ex

Bug#949310: buster-pu: package gnutls28/3.6.7-4+deb10u1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: buster
User: release.debian@packages.debian.org
Usertags: pu

Hello,

there is a regression in gnutls/buster compared to stretch. It fails to
parse certificates using Registered ID in Subject Alternative Name.

See upstream report https://gitlab.com/gnutls/gnutls/issues/905 for more
details.

I would like to fix this in pu, by pulling the fix from GnuTLS 3.6.9.
The respective upstream change also adds a testcase and therefore
adds/modifies binaries. The proposed Debian changes are not
representable as debdiff, I am attaching git-format-patch diff instead.

cu Andreas
From de3d573242195eddab914709584242610b2e2762 Mon Sep 17 00:00:00 2001
From: Andreas Metzler 
Date: Sun, 19 Jan 2020 18:00:12 +0100
Subject: [PATCH] Fix parsing of certificates using RegisteredID Closes:
 #949293

---
 debian/binary/cert10.der  | Bin 0 -> 571 bytes
 debian/binary/cert5.der   | Bin 0 -> 414 bytes
 debian/changelog  |   6 +
 ...ralname-registeredID-from-RFC-5280-i.patch | 242 ++
 debian/patches/series |   1 +
 debian/rules  |   8 +
 debian/source/include-binaries|   2 +
 7 files changed, 259 insertions(+)
 create mode 100644 debian/binary/cert10.der
 create mode 100644 debian/binary/cert5.der
 create mode 100644 debian/patches/41_rel3.6.9_01-Support-for-Generalname-registeredID-from-RFC-5280-i.patch

diff --git a/debian/binary/cert10.der b/debian/binary/cert10.der
new file mode 100644
index ..07ab16d3eec034bd14cd94dd0174a2a76c768918
GIT binary patch
literal 571
zcmXqLVlp>qV!XS6nTe5!i7~~1i;Y98~}r=h5UFdK6y3o{Rod$6ygLP%`WXB7yE<
z2fL4n5$aH8Ms{W=1{U9cSH5JrPuwpy_1pr3D$^|zjMJuCRBw;2RdL_8Rbf7h>pH)<
zAeoC69oOR@)U-AzX+7fIwMMr5Y?*=QWGCtCmWvy28Z=%rkOx{StIQ%{Al4xA)v;*r
z&#

Bug#947365: transition: libvigraimpex

[Andreas Metzler]

On 2020-01-03 Paul Gevers  wrote:
> On 31-12-2019 18:26, Andreas Metzler wrote:
[...] 
> > Afaict the involved packages should propagate to testing in 3 days, when
> > enblend-enfuse is old enough. I have commited the fix. [1]

> Unfortunately libvigraimpex is (hopefully only temporarily) blocked by a
> bug we had in piuparts and the (automatically) rescheduling of the
> failed tests takes some days [1]. So, maybe it would be even faster to
> upload the fix now.

> Paul

> [1] I don't want to manually test all the piuparts failures as there are
> too many, and I don't want to blindly ignore those results.

Hello Paul,

I have just dput-ed. Thanks for the monitoring.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#947365: transition: libvigraimpex

[Andreas Metzler]

On 2019-12-31 Sebastiaan Couwenberg  wrote:
> On 12/31/19 4:20 PM, Andreas Metzler wrote:
>> as Bas correctly diagnoses I am not currently building for all supported
>> versions but only for the default one because it is not trivial but
>> requires some work. Looking at python policy I think that is acceptable
>> but not perfect.

>> Is my reading of polic correct?

> AFAIK the python policy doesn't document the build dependencies.

> doko filed bugs for at least one of my packages that had python3-all-dev
> in B-D but only built for the default interpreter with the request to
> change the B-D or build for all supported versions, so either option is
> fine.

Hello,

Thanks for clearing this up. I was just unsure whether not building
for all suppported versions was acceptable.

>> Shall I make a timely upload fixing the
>> build-dependency or can I wait for propagation of vigra packages to
>> testing?

> I would commit the change now, and upload it after the testing migration
> unless there are other blockers that hold up the migration for more than
> 5 days, then I would upload it now.

Afaict the involved packages should propagate to testing in 3 days, when
enblend-enfuse is old enough. I have commited the fix. [1]

cu Andreas

[1] 
https://salsa.debian.org/ametzler/libvigraimpex/commit/c5a8c27cc018c8968036b73c39d0d02d0f229320

Bug#947365: transition: libvigraimpex

[Andreas Metzler]

On 2019-12-31 Sebastiaan Couwenberg  wrote:
> On 12/30/19 9:48 PM, Paul Gevers wrote:
[...]
>> libvigraimpex is also part of the pseudo python3.8 transition [1], but
>> it is still red. This probably means that you are not correctly building
>> Python3 modules for all supported Python3 versions. Can you please check?

> Looks like it shouldn't build depend on python3-all-dev since the build
> systems only uses the default interpreter.

>  sed -i 's/python3-all-dev/python3-dev/g' debian/control

Hello,

as Bas correctly diagnoses I am not currently building for all supported
versions but only for the default one because it is not trivial but
requires some work. Looking at python policy I think that is acceptable
but not perfect.

Is my reading of polic correct? Shall I make a timely upload fixing the
build-dependency or can I wait for propagation of vigra packages to
testing?

Sorry for the inconvenience.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#947365: transition: libvigraimpex

[Andreas Metzler]

On 2019-12-26 Paul Gevers  wrote:
> On 25-12-2019 19:29, Andreas Metzler wrote:
>> libvigraimpex is marked for autoremoval because of the python2 removal.
>> This is fixed in experimental, the new version features a soname bump.
[...]
> Normally we don't want python 2 removal package uploads and transitions
> mixed, but it seems that python-vigra doesn't have reverse dependencies.
> Please go ahead in unstable.

Thank you, uploaded.

cu Andreas

Bug#947365: transition: libvigraimpex

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

Hello,

libvigraimpex is marked for autoremoval because of the python2 removal.
This is fixed in experimental, the new version features a soname bump.
this should be a small scale transition. I have successfully [1] rebuilt
3depict enblend-enfuse hugin luminance-hdr saga against -3 in
experimental.


Ben file:

title = "libvigraimpex";
is_affected = .depends ~ "libvigraimpex6" | .depends ~ "libvigraimpex11";
is_good = .depends ~ "libvigraimpex11";
is_bad = .depends ~ "libvigraimpex6";

cu Andreas

[1] 3depic is broken by https://bugs.debian.org/947364

Bug#946644: nmu: sa-exim_4.2.1-18

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: binnmu

nmu sa-exim_4.2.1-18 . ANY . unstable . -m "Rebuild against 
exim4-localscanapi-3.1 (exim 4.93)"

Hello,

please binNMU sa-exim to make it installable again.

exim 4.93 broke the local_scan exim and therefore the localscanapi
version was bumped. ZThis bump happened in RC5 after sa-exim had
already been updated for the API changes.

cu Andreas

Bug#945925: buster-pu: package gnutls28/3.6.7-4+deb10u1

[Andreas Metzler]

format)
--
Build-Ids: [-d328298de34135fca5f236357f2f2dd56cb109f3-] {+b17a60f0701c7de3d7e5e921305846b5efbc3c91+}
Depends: libgnutls-dane0 (= [-3.6.7-4)-] {+3.6.7-4+deb10u1)+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutls-openssl27: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libc6 (>= 2.14)
Installed-Size: [-372-] {+373+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutls-openssl27-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-fe4c3c0c38af44779c38ae5d1e187b6250f7afe0-] {+5e61e31c2ae39982eeb14ae1c8f66aff43e1083a+}
Depends: libgnutls-openssl27 (= [-3.6.7-4)-] {+3.6.7-4+deb10u1)+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutls28-dev: lines which differ (wdiff format)
---
Depends: libc6-dev | libc-dev, libgnutls-dane0 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libgnutls-openssl27 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libgnutls30 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libgnutlsxx28 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libidn2-dev, libp11-kit-dev (>= 0.23.10), libtasn1-6-dev, nettle-dev (>= 3.4.1~rc1)
Installed-Size: [-4313-] {+4314+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutls30: lines which differ (wdiff format)
---
Installed-Size: [-2643-] {+2644+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutls30-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-4d66d28cd2e7537e1e1d2905595b260226b22ad2-] {+1ca9574531f2bffce01464c8a654b2e0c2ed894b+}
Depends: libgnutls30 (= [-3.6.7-4)-] {+3.6.7-4+deb10u1)+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutlsxx28: lines which differ (wdiff format)
-
Depends: libgnutls30 (= [-3.6.7-4),-] {+3.6.7-4+deb10u1),+} libc6 (>= 2.14), libgcc1 (>= 1:3.0), libstdc++6 (>= 5)
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}

Control files of package libgnutlsxx28-dbgsym: lines which differ (wdiff format)

Build-Ids: [-d752158b357b5875ebc8680001b57a886b94a1a4-] {+b8bd0e5aecb48c352850674891129476d08d016a+}
Depends: libgnutlsxx28 (= [-3.6.7-4)-] {+3.6.7-4+deb10u1)+}
Version: [-3.6.7-4-] {+3.6.7-4+deb10u1+}



diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog	2019-06-12 19:21:23.0 +0200
+++ gnutls28-3.6.7/debian/changelog	2019-11-30 13:41:59.0 +0100
@@ -1,3 +1,11 @@
+gnutls28 (3.6.7-4+deb10u1) buster; urgency=medium
+
+  * 42_rel3.6.10_01-gnutls_epoch_set_keys-do-not-forbid-random-padding.patch
+from 3.6.10: Fix interop problems with gnutls 2.x. Closes: #933538
+(Thanks, Hanno Stock!)
+
+ -- Andreas Metzler   Sat, 30 Nov 2019 13:41:59 +0100
+
 gnutls28 (3.6.7-4) unstable; urgency=medium
 
   * Cherry-pick important bug-fixes from 3.6.8:
diff -Nru gnutls28-3.6.7/debian/patches/42_rel3.6.10_01-gnutls_epoch_set_keys-do-not-forbid-random-padding.patch gnutls28-3.6.7/debian/patches/42_rel3.6.10_01-gnutls_epoch_set_keys-do-not-forbid-random-padding.patch
--- gnutls28-3.6.7/debian/patches/42_rel3.6.10_01-gnutls_epoch_set_keys-do-not-forbid-random-padding.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.6.7/debian/patches/42_rel3.6.10_01-gnutls_epoch_set_keys-do-not-forbid-random-padding.patch	2019-11-30 13:41:59.0 +0100
@@ -0,0 +1,63 @@
+From daa49b9e455d262a1a2bc1b641e72dc004e2cb3e Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos 
+Date: Sat, 3 Aug 2019 21:51:58 +0200
+Subject: [PATCH] _gnutls_epoch_set_keys: do not forbid random padding in
+ TLS1.x CBC ciphersuites
+
+Since some point in 3.6.x we updated the calculation of maximum record size,
+however that did not include the possibility of random record padding available
+for CBC ciphersuites which exceeds the maximum. This commit allows for larger
+sizes for these ciphersuites to account for random padding as applied by
+gnutls 2.12.x.
+
+Resolves: #811
+
+Signed-off-by: Nikos Mavrogiannopoulos 
+---
+ NEWS   |  4 
+ lib/constate.c | 11 +--
+ lib/record.c   |  4 ++--
+ 3 files changed, 15 insertions(+), 4 deletions(-)
+
+diff --git a/lib/constate.c b/lib/constate.c
+index 51a4eca30..4c6ca0fd0 100644
+--- a/lib/constate.c
 b/lib/constate.c
+@@ -707,10 +707,17 @@ int _gnutls_epoch_set_keys(gnutls_session_t session, uint16_t epoch, hs_stage_t
+ 			return gnutls_assert_val(ret);
+ 	}
+ 
+-	if (ver->tls13_sem) {
++	/* The TLS1.3 li

Bug#939595: release.debian.org: Please bump urgency of exim4 4.92.1-3 to criticial

[Andreas Metzler]

Package: release.debian.org
Severity: normal

Hello,

In hindsight I have chosen the wrong urgency for exim4 4.92.1-3, a
remote root command execution should migrate as fast as possible. Could
please bump the urgency/shorten the migration period?

urgent exim4/4.92.1-3

TIA, cu Andreas

-- System Information:
Debian Release: bullseye/sid
  APT prefers testing-debug
  APT policy: (500, 'testing-debug'), (500, 'testing'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 5.2.0-2-amd64 (SMP w/4 CPU cores)
Locale: LANG=de_AT.UTF-8, LC_CTYPE=de_AT.UTF-8 (charmap=UTF-8), 
LANGUAGE=de_AT.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#930490: unblock: exim4/4.92-8

[Andreas Metzler]

set to a directory instead of a file
+  exim/GnuTLS would still send out the list of accepted certificates,
+  This did not match documented behavior.
++ 75_13-Use-dsn_from-for-success-DSN-messages.-Bug-2404.patch
+  The dsn_from option was not used for DSN success messages.
+  * Pulled from upstream GIT master:
++ 75_14-Fix-smtp-response-timeout.patch
+  Fix the timeout on smtp response to apply to the whole response instead
+  of resetting for every byte received.
++ 75_15-Fix-detection-of-32b-platform-at-build-time.-Bug-240.patch
+  https://bugs.exim.org/show_bug.cgi?id=2405
+  ${eval } was broken on 32bit archs.
+
+ -- Andreas Metzler   Sat, 08 Jun 2019 17:37:43 +0200
+
 exim4 (4.92-7) unstable; urgency=medium
 
   * Upload to unstable.
diff -Nru 
exim4-4.92/debian/patches/75_11-GnuTLS-fix-tls_out_ocsp-under-hosts_request_ocsp.patch
 
exim4-4.92/debian/patches/75_11-GnuTLS-fix-tls_out_ocsp-under-hosts_request_ocsp.patch
--- 
exim4-4.92/debian/patches/75_11-GnuTLS-fix-tls_out_ocsp-under-hosts_request_ocsp.patch
  1970-01-01 01:00:00.0 +0100
+++ 
exim4-4.92/debian/patches/75_11-GnuTLS-fix-tls_out_ocsp-under-hosts_request_ocsp.patch
  2019-06-08 13:49:59.0 +0200
@@ -0,0 +1,54 @@
+From 5e64b73ef7cdaf20b998b3345a588b462fd30bfb Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Tue, 7 May 2019 22:55:41 +0100
+Subject: [PATCH] GnuTLS: fix $tls_out_ocsp under hosts_request_ocsp
+
+(cherry picked from commit 7a501c874f028f689c44999ab05bb0d39da46941)
+---
+ doc/ChangeLog |  3 +++
+ src/tls-gnu.c | 12 
+ test/log/5651 |  2 +-
+ test/log/5730 |  8 
+ 4 files changed, 16 insertions(+), 9 deletions(-)
+
+--- a/doc/ChangeLog
 b/doc/ChangeLog
+@@ -39,6 +39,9 @@ JH/11 Harden plaintext authenticator aga
+   library routine (usually a crash).  Found by "zerons".
+ 
+ 
++JH/18 GnuTLS: fix $tls_out_ocsp under hosts_request_ocsp. Previously the
++  verification result was not updated unless hosts_require_ocsp applied.
++
+ 
+ Exim version 4.92
+ -
+--- a/src/tls-gnu.c
 b/src/tls-gnu.c
+@@ -2450,7 +2450,7 @@ if (!verify_certificate(state, errstr))
+   }
+ 
+ #ifndef DISABLE_OCSP
+-if (require_ocsp)
++if (request_ocsp)
+   {
+   DEBUG(D_tls)
+ {
+@@ -2474,10 +2474,14 @@ if (require_ocsp)
+ {
+ tlsp->ocsp = OCSP_FAILED;
+ tls_error(US"certificate status check failed", NULL, state->host, errstr);
+-return NULL;
++if (require_ocsp)
++  return FALSE;
++}
++  else
++{
++DEBUG(D_tls) debug_printf("Passed OCSP checking\n");
++tlsp->ocsp = OCSP_VFIED;
+ }
+-  DEBUG(D_tls) debug_printf("Passed OCSP checking\n");
+-  tlsp->ocsp = OCSP_VFIED;
+   }
+ #endif
+ 
diff -Nru 
exim4-4.92/debian/patches/75_12-GnuTLS-fix-the-advertising-of-acceptable-certs-by-th.patch
 
exim4-4.92/debian/patches/75_12-GnuTLS-fix-the-advertising-of-acceptable-certs-by-th.patch
--- 
exim4-4.92/debian/patches/75_12-GnuTLS-fix-the-advertising-of-acceptable-certs-by-th.patch
  1970-01-01 01:00:00.0 +0100
+++ 
exim4-4.92/debian/patches/75_12-GnuTLS-fix-the-advertising-of-acceptable-certs-by-th.patch
  2019-06-08 13:50:55.0 +0200
@@ -0,0 +1,42 @@
+From 44893ba5249c6c6d5a0d62a1cc57ba3fbf7185b4 Mon Sep 17 00:00:00 2001
+From: Jeremy Harris 
+Date: Sun, 19 May 2019 12:12:36 +0100
+Subject: [PATCH 1/2] GnuTLS: fix the advertising of acceptable certs by the
+ server.  Bug 2389
+
+(cherry picked from commit 12d95aa62042377fc9f603245a17a43142972447)
+---
+ doc/ChangeLog | 4 
+ src/tls-gnu.c | 8 
+ 2 files changed, 12 insertions(+)
+
+--- a/doc/ChangeLog
 b/doc/ChangeLog
+@@ -42,6 +42,10 @@ JH/11 Harden plaintext authenticator aga
+ JH/18 GnuTLS: fix $tls_out_ocsp under hosts_request_ocsp. Previously the
+   verification result was not updated unless hosts_require_ocsp applied.
+ 
++JH/20 Bug 2389: fix server advertising of usable certificates, under GnuTLS in
++  directory-of-certs mode.  Previously they were advertised despite the
++  documentation.
++
+ 
+ Exim version 4.92
+ -
+--- a/src/tls-gnu.c
 b/src/tls-gnu.c
+@@ -1133,6 +1133,14 @@ else
+ #endif
+ gnutls_certificate_set_x509_trust_file(state->x509_cred,
+   CS state->exp_tls_verify_certificates, GNUTLS_X509_FMT_PEM);
++
++#ifdef SUPPORT_CA_DIR
++  /* Mimic the behaviour with OpenSSL of not advertising a usable-cert list
++  when using the directory-of-certs config model. */
++
++  if ((statbuf.st_mode & S_IFMT) == S_IFDIR)
++gnutls_certificate_send_x509_rdn_sequence(state->session, 1);
++#endif
+   }
+ 
+ if (cert_count < 0)
diff -Nru 
exim4-4.92/debian/patches/75_13-Use-dsn_from-for-success-DSN-messages.-Bug-2404.patch
 
exim4-4.92/debian/patches/75_13-Use-dsn_from-for-success-DSN-messages.-Bug-2404.patch
--- 
exim4-4.92/debian/patches/75_13-Use-dsn_from-for-success-DSN-messages.-Bug-2404

Bug#930491: unblock: gnutls28/3.6.7-4

[Andreas Metzler]

[-3.6.7-3)-] {+3.6.7-4)+}
Version: [-3.6.7-3-] {+3.6.7-4+}

Control files of package libgnutls28-dev: lines which differ (wdiff format)
---
Depends: libc6-dev | libc-dev, libgnutls-dane0 (= [-3.6.7-3),-] {+3.6.7-4),+} 
libgnutls-openssl27 (= [-3.6.7-3),-] {+3.6.7-4),+} libgnutls30 (= [-3.6.7-3),-] 
{+3.6.7-4),+} libgnutlsxx28 (= [-3.6.7-3),-] {+3.6.7-4),+} libidn2-dev, 
libp11-kit-dev (>= 0.23.10), libtasn1-6-dev, nettle-dev (>= 3.4.1~rc1)
Installed-Size: [-4312-] {+4313+}
Version: [-3.6.7-3-] {+3.6.7-4+}

Control files of package libgnutls30: lines which differ (wdiff format)
---
Version: [-3.6.7-3-] {+3.6.7-4+}

Control files of package libgnutls30-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-febecd51bb621afd4a8f0352f55d6c2ed96df57a-] 
{+4d66d28cd2e7537e1e1d2905595b260226b22ad2+}
Depends: libgnutls30 (= [-3.6.7-3)-] {+3.6.7-4)+}
Installed-Size: [-4058-] {+4061+}
Version: [-3.6.7-3-] {+3.6.7-4+}

Control files of package libgnutlsxx28: lines which differ (wdiff format)
-
Depends: libgnutls30 (= [-3.6.7-3),-] {+3.6.7-4),+} libc6 (>= 2.14), libgcc1 
(>= 1:3.0), libstdc++6 (>= 5)
Version: [-3.6.7-3-] {+3.6.7-4+}

Control files of package libgnutlsxx28-dbgsym: lines which differ (wdiff format)

Depends: libgnutlsxx28 (= [-3.6.7-3)-] {+3.6.7-4)+}
Version: [-3.6.7-3-] {+3.6.7-4+}
diff -Nru gnutls28-3.6.7/debian/changelog gnutls28-3.6.7/debian/changelog
--- gnutls28-3.6.7/debian/changelog 2019-05-19 10:48:52.0 +0200
+++ gnutls28-3.6.7/debian/changelog 2019-06-12 19:21:23.0 +0200
@@ -1,3 +1,28 @@
+gnutls28 (3.6.7-4) unstable; urgency=medium
+
+  * Cherry-pick important bug-fixes from 3.6.8:
++ 40_rel3.6.8_01-gnutls_srp_entry_free-follow-consistent-behavior-in.patch
+  The gnutls_srp_set_server_credentials_function can be used with the 8192
+  parameters as well.
+  https://gitlab.com/gnutls/gnutls/issues/761
++ 40_rel3.6.8_05-lib-nettle-fix-carry-flag-in-Streebog-code.patch
+  Fix calculation of Streebog digests (incorrect carry operation in
+  512 bit addition).
++ 40_rel3.6.8_10-ext-record_size_limit-distinguish-sending-and-receiv.patch
+  Fix compatibility of GnuTLS 3.6.[456] server with GnuTLS 3.6.7 client.
+  Closes: #929907
++ 40_rel3.6.8_15-Apply-STD3-ASCII-rules-in-gnutls_idna_map.patch
+  Apply STD3 ASCII rules in gnutls_idna_map() to prevent hostname/domain
+  crafting via IDNA conversion.
+  https://gitlab.com/gnutls/gnutls/issues/720
++ 40_rel3.6.8_20-pubkey-remove-deprecated-TLS1_RSA-flag-check.patch
+  Fixed bug preventing the use of gnutls_pubkey_verify_data2() and
+  gnutls_pubkey_verify_hash2() with the GNUTLS_VERIFY_DISABLE_CA_SIGN
+  flag.
+  https://gitlab.com/gnutls/gnutls/issues/754
+
+ -- Andreas Metzler   Wed, 12 Jun 2019 19:21:23 +0200
+
 gnutls28 (3.6.7-3) unstable; urgency=medium
 
   * Revert debhelper upgrade, use DH 10.
diff -Nru 
gnutls28-3.6.7/debian/patches/40_rel3.6.8_01-gnutls_srp_entry_free-follow-consistent-behavior-in.patch
 
gnutls28-3.6.7/debian/patches/40_rel3.6.8_01-gnutls_srp_entry_free-follow-consistent-behavior-in.patch
--- 
gnutls28-3.6.7/debian/patches/40_rel3.6.8_01-gnutls_srp_entry_free-follow-consistent-behavior-in.patch
  1970-01-01 01:00:00.0 +0100
+++ 
gnutls28-3.6.7/debian/patches/40_rel3.6.8_01-gnutls_srp_entry_free-follow-consistent-behavior-in.patch
  2019-06-12 19:21:15.0 +0200
@@ -0,0 +1,65 @@
+From 0bdca5d51f203cf414d645e75ac197e3fadfadc8 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos 
+Date: Fri, 10 May 2019 06:30:12 +0200
+Subject: [PATCH] _gnutls_srp_entry_free: follow consistent behavior in freeing
+ data
+
+_gnutls_srp_entry_free would previously not free any parameters that
+were known to gnutls to account for documented behavior of
+gnutls_srp_set_server_credentials_function(). This was not updated
+when the newly added 8192 parameter was added to the library.
+
+This introduces a safety check for generator parameters, even though
+in practice they are the same pointer.
+
+Resolves: #761
+
+Signed-off-by: Nikos Mavrogiannopoulos 
+---
+ NEWS  |  3 +++
+ lib/auth/srp_passwd.c | 12 
+ 2 files changed, 11 insertions(+), 4 deletions(-)
+
+--- a/NEWS
 b/NEWS
+@@ -47,6 +47,9 @@ See the end for copying conditions.
+ 
+ ** gnutls-cli: Added option --logfile to redirect informational messages 
output.
+ 
++** libgnutls: the gnutls_srp_set_server_credentials_function can be used
++   with the 8192 parameters as well (#995).
++
+ ** API and ABI modifications:
+ No changes since last version.
+ 
+--- a/lib/auth/srp_

Bug#930490: unblock: exim4/4.92-8

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package exim4. This upload pulls 5 patches from upstream
GIT:
+ 75_11-GnuTLS-fix-tls_out_ocsp-under-hosts_request_ocsp.patch
  Fix expansion of $tls_out_ocsp under hosts_request_ocsp.
+ 75_12-GnuTLS-fix-the-advertising-of-acceptable-certs-by-th.patch
  When tls_verify_certificates was set to a directory instead of a file
  exim/GnuTLS would still send out the list of accepted certificates,
  This did not match documented behavior.
+ 75_13-Use-dsn_from-for-success-DSN-messages.-Bug-2404.patch
  The dsn_from option was not used for DSN success messages.
+ 75_14-Fix-smtp-response-timeout.patch
  Fix the timeout on smtp response to apply to the whole response instead
  of resetting for every byte received.
+ 75_15-Fix-detection-of-32b-platform-at-build-time.-Bug-240.patch
  https://bugs.exim.org/show_bug.cgi?id=2405
  ${eval } was broken on 32bit archs.

unblock exim4/4.92-8

Thanks, cu Andreas

Bug#926412: unblock: gnutls28/3.6.7-2

[Andreas Metzler]

On 2019-05-20 Paul Gevers  wrote:
> On 19-05-2019 10:33, Andreas Metzler wrote:
>> I probably could try to pick the CVE related changes and other important
>> bug-fixes, however I do not think it is the right choice. The changes
>> will be smaller but the risk of breakage is higher.

> Can you explain why do you believe that?

>> Also 3.6.7 now has
>> been tested in sid for almost two months now. 

> Ack.

Hello Paul,

well, apart from the two CVE fixes there are many bugfixes in this
release that we probably want, e.g.
https://gitlab.com/gnutls/gnutls/issues/690
https://gitlab.com/gnutls/gnutls/issues/689
https://gitlab.com/gnutls/gnutls/issues/713
https://gitlab.com/gnutls/gnutls/issues/698
etc.

Most of these are related to TLS 1.3. - They might not show up as bug
reports now because it TLS1.3 is not that common yet but will propably
cause issues later in buster's lifetime. And the more fixes there the
more error-prone complicated cherry-picking s going to be.

>>> You bumped the debhelper compat level. That isn't a change we find
>>> acceptable during the freeze.
>> 
>> I will immediately revert this if it helps.

> I don't have enough experience yet with reviewing unblocks, that I feel
> comfortable reviewing and unblocking the current package, so if your
> insisting on the whole, somebody else will have to do the review. I am
> sure this revert will be a requirement though.

The revert has been in sid for a week now.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#926412: unblock: gnutls28/3.6.7-2

[Andreas Metzler]

On 2019-05-18 Paul Gevers  wrote:
[gnutls]
> Is it reasonably possible to split of the CVE changes and patch the
> version currently in testing? That would be much more comfortable for
> us. Either by reverting the new upstream version with e.g. an +really
> version number, or, but less preferred by us, via an upload to
> testing-proposed-updates.

Hello Paul,

I probably could try to pick the CVE related changes and other important
bug-fixes, however I do not think it is the right choice. The changes
will be smaller but the risk of breakage is higher. Also 3.6.7 now has
been tested in sid for almost two months now. 

> You bumped the debhelper compat level. That isn't a change we find
> acceptable during the freeze.

I will immediately revert this if it helps.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#926412: unblock: gnutls28/3.6.7-2

[Andreas Metzler]

On 2019-04-04 Andreas Metzler  wrote:
[...]
> This is a upstream bugfix release featuring two security fixes

> + Fixes a memory corruption (double free) vulnerability in the
>   certificate verification API.
>   https://gitlab.com/gnutls/gnutls/issues/694 CVE-2019-3829
>   GNUTLS-SA-2019-03-27
> + Fixes an invalid pointer access via malformed TLS1.3 async messages;
>   https://gitlab.com/gnutls/gnutls/issues/704 CVE-2019-3836
>   GNUTLS-SA-2019-03-27
[...]

Ping?

Bug#926878: unblock: exim4/4.92-5

[Andreas Metzler]

On 2019-04-11 Paul Gevers  wrote:
> Control: tags -1 moreinfo

> Hi Andreas,

> On 11-04-2019 19:51, Andreas Metzler wrote:
>> The second notable change is related to sa-exim. Exim in Debian was
>> patched to allow dlopening a localscan() module. The single consumer of
>> this patch in Debian is sa-exim. (The patch also originates there.)

>> The patch in Debian has been nonfunctional in unstable for quite some
>> time (4.92~RC2-1/experimental/18 Dec, 4.92~RC3-1 unstable/26 Dec and
>> buster/03 Jan). The issue only popped up end of March on the upstream
>> user support ML.

>> Looking at the state of sa-exim (dead upstream since 2006 and buggy: 
>> https://lists.exim.org/lurker/message/20180726.113354.6d03efde.en.html
>> #879687) we have decided stop patching exim, which resulted in 4.92-5,
>> which
>> - improves the example/docs for content-scanning in exim without sa-exim
>> - drops the abovementioned patch and the virtual Provides for
>>   exim4-localscanapi-2.0 and also drops the exim-dev packages (only
>>   needed for sa-exim). Exim now also Conflicts with sa-exim.

> I am probably missing something, but as far as I see it, your packages
> can't migrate to testing/buster because it would make sa-exim
> uninstallable.

uninstallable and unbuildable.

> If I am right, please coordinate with the maintainer of
> sa-exim (in CC). At least at this moment they should agree that it is
> alright to remove sa-exim from buster. I am not seeing any serious bugs
> reported against sa-exim so they may not be aware of the issue.

I had X-Debbugs-Cc'ed sa-exim@pdo on this report. I will also open a rc
bug suggesting sa-exim removal to hash this out properly.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#926878: unblock: exim4/4.92-5

[Andreas Metzler]

a65884d62d6506ce390b4f07 5955fdc7b64bc2f31b1e0b63c762a57924c2516e 9b6cfa23511aa8ae2305e45f556cd5238b07f495 bb23e5a1a9f351c2a608d482dfc1e00d9998c629 bc986da4b151ecfa52558aa9c20d03614d31dd25+}
Depends: exim4-base (= [-4.92-2)-] {+4.92-5)+}
Version: [-4.92-2-] {+4.92-5+}

Control files of package exim4-config: lines which differ (wdiff format)

Installed-Size: [-983-] {+985+}
Version: [-4.92-2-] {+4.92-5+}

Control files of package exim4-daemon-heavy: lines which differ (wdiff format)
--
Conflicts: [-mail-transport-agent-] {+mail-transport-agent, sa-exim+}
Installed-Size: [-1477-] {+1537+}
Provides: [-exim4-localscanapi-2.0,-] mail-transport-agent
Version: [-4.92-2-] {+4.92-5+}

Control files of package exim4-daemon-heavy-dbgsym: lines which differ (wdiff format)
-
Build-Ids: [-50c2969f4b54bc47c33c513e27a89cd4a09d728d-] {+bd894614600fc329441d05ceb08017719b489417+}
Depends: exim4-daemon-heavy (= [-4.92-2)-] {+4.92-5)+}
Installed-Size: [-2646-] {+2631+}
Version: [-4.92-2-] {+4.92-5+}

Control files of package exim4-daemon-light: lines which differ (wdiff format)
--
Conflicts: [-mail-transport-agent-] {+mail-transport-agent, sa-exim+}
Installed-Size: [-1332-] {+1324+}
Provides: default-mta, [-exim4-localscanapi-2.0,-] mail-transport-agent
Version: [-4.92-2-] {+4.92-5+}

Control files of package exim4-daemon-light-dbgsym: lines which differ (wdiff format)
-
Build-Ids: [-51279c0f518a9e2a849c64a89ff8eaadcabe26fa-] {+caa4ade19a8e042ebf7f9f22782142cbd56bcd2b+}
Depends: exim4-daemon-light (= [-4.92-2)-] {+4.92-5)+}
Installed-Size: [-2260-] {+2247+}
Version: [-4.92-2-] {+4.92-5+}

Control files of package eximon4: lines which differ (wdiff format)
---
Installed-Size: [-212-] {+216+}
Version: [-4.92-2-] {+4.92-5+}

Control files of package eximon4-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-a31149847f6ae982b262e6aec59d3afa2e9ae841-] {+5ef1dbf7d44b659418b55dd4a173cda74ecad278+}
Depends: eximon4 (= [-4.92-2)-] {+4.92-5)+}
Version: [-4.92-2-] {+4.92-5+}



diff -Nru exim4-4.92/debian/changelog exim4-4.92/debian/changelog
--- exim4-4.92/debian/changelog	2019-02-20 19:23:11.0 +0100
+++ exim4-4.92/debian/changelog	2019-04-07 13:39:31.0 +0200
@@ -1,3 +1,33 @@
+exim4 (4.92-5) unstable; urgency=medium
+
+  * Improved spam-scanning example with accompaning information in
+README.Debian. Explicitly warn about adding the default SpamAssassin
+report in a header, which Closes: #774553
+  * Drop 90_localscan_dlopen.dpatch. (It has been non-functional for a couple
+of months.) Closes: #925982 Add a Conflicts for sa-exim, which relied on
+the (working) version of the patch. Drop exim4-dev package. Add a NEWS
+entry for this change.
+
+ -- Andreas Metzler   Sun, 07 Apr 2019 13:39:31 +0200
+
+exim4 (4.92-4) unstable; urgency=medium
+
+  * Another patch from exim-4.92+fixes branch:
+75_10-Harden-plaintext-authenticator.patch
+
+ -- Andreas Metzler   Fri, 22 Mar 2019 07:15:20 +0100
+
+exim4 (4.92-3) unstable; urgency=medium
+
+  * Pull fixes from exim-4.92+fixes branch.
++ 75_05-Fix-expansions-for-RFC-822-addresses-having-comments.patch
++ 75_06-Docs-Add-note-on-lsearch-for-IPv4-mapped-IPv6-addres.patch
++ 75_07-Fix-crash-from-SRV-lookup-hitting-a-CNAME.patch
++ 75_08-Logging-fix-initial-listening-on-log-line.patch
++ 75_09-OpenSSL-Fix-aggregation-of-messages.patch
+
+ -- Andreas Metzler   Wed, 20 Mar 2019 17:01:29 +0100
+
 exim4 (4.92-2) unstable; urgency=medium
 
   * Upload to unstable.
diff -Nru exim4-4.92/debian/control exim4-4.92/debian/control
--- exim4-4.92/debian/control	2019-02-17 13:13:18.0 +0100
+++ exim4-4.92/debian/control	2019-04-07 13:39:31.0 +0200
@@ -138,10 +138,9 @@
 Architecture: any
 Priority: optional
 Provides:
- exim4-localscanapi-2.0,
  mail-transport-agent,
  ${dist:Provides:exim4-daemon-light}
-Conflicts: mail-transport-agent
+Conflicts: mail-transport-agent, sa-exim
 Replaces: exim4-base (<= 4.61-1), mail-transport-agent
 Depends:
  exim4-base (>= ${Upstream-Version}),
@@ -202,8 +201,8 @@
 Package: exim4-daemon-heavy
 Architecture: any
 Priority: optional
-Provides: exim4-localscanapi-2.0, mail-transport-agent
-Conflicts: mail-transport-agent
+Provides: mail-transport-agent
+Conflicts: mail-transport-agent, sa-exim
 Replaces: exim4-base (<= 4.61-1), mail-transport-agent
 Depends:
  exim4-base (>= ${Upstream-Version}),
@@ -238,8 +237,8 @@
 #Package: exim4-daemon-custom
 #Architecture: any
 #Priority: optional
-#Provides: exim4-lo

Bug#926412: unblock: gnutls28/3.6.7-2

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package gnutls28.

This is a upstream bugfix release featuring two security fixes

+ Fixes a memory corruption (double free) vulnerability in the
  certificate verification API.
  https://gitlab.com/gnutls/gnutls/issues/694 CVE-2019-3829
  GNUTLS-SA-2019-03-27
+ Fixes an invalid pointer access via malformed TLS1.3 async messages;
  https://gitlab.com/gnutls/gnutls/issues/704 CVE-2019-3836
  GNUTLS-SA-2019-03-27

One of these is fixed by a hardening measure (gnutls_free() will
automatically set the free'd pointer to NULL.) It also unbreaks
vlc (#922879) and has some TLS1.3 related changes.

The straight debdiff is huge, because of a) usual release updates of
autogenerated files and b) because it includes a global
's/http:/https:/'. Stripped down debdiff is attached.

unblock gnutls28/3.6.7-2

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


smaller.debdiff.diff.xz
Description: application/xz


signature.asc
Description: PGP signature

Bug#910445: stretch-pu: package gnutls28/3.5.8-5+deb9u4

[Andreas Metzler]

+deb9u4+}

Control files of package libgnutls-dane0: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} libc6 (>= 
2.14), libunbound2 (>= 1.4.1)
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls-dane0-dbgsym: lines which differ (wdiff 
format)
--
Build-Ids: [-25228bbeb1c692f8764099a856ab8c9463f7c325-] 
{+1c399494f95f5e9ff28fcbd0243e96639fad69d3+}
Depends: libgnutls-dane0 (= [-3.5.8-5+deb9u3)-] {+3.5.8-5+deb9u4)+}
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls-openssl27: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} libc6 (>= 
2.14)
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls-openssl27-dbgsym: lines which differ (wdiff 
format)
--
Build-Ids: [-baf0016a0105eb9eb689bd33997207d4a704386d-] 
{+51a6d9549543590e69584a2dd9df4e919cd62918+}
Depends: libgnutls-openssl27 (= [-3.5.8-5+deb9u3)-] {+3.5.8-5+deb9u4)+}
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls28-dev: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} 
libgnutls-openssl27 (= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} libgnutlsxx28 
(= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} libgnutls-dane0 (= 
[-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} nettle-dev, libc6-dev | libc-dev, 
zlib1g-dev, libtasn1-6-dev, libp11-kit-dev, libidn11-dev (>= 1.31)
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls30: lines which differ (wdiff format)
---
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutls30-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-07a8f58a7e4e32a36feee7511f728d5896439b13-] 
{+1c1bc93c559cfe2ebd1b5676fa4b355118edf38e+}
Depends: libgnutls30 (= [-3.5.8-5+deb9u3)-] {+3.5.8-5+deb9u4)+}
Installed-Size: [-2880-] {+2882+}
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutlsxx28: lines which differ (wdiff format)
-
Depends: libgnutls30 (= [-3.5.8-5+deb9u3),-] {+3.5.8-5+deb9u4),+} libc6 (>= 
2.4), libgcc1 (>= 1:3.0), libstdc++6 (>= 5)
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

Control files of package libgnutlsxx28-dbgsym: lines which differ (wdiff format)

Build-Ids: [-a8a2ad066f20b10398a4047b4a5ac2032fdcc3d7-] 
{+f443a08baf0b78f1286c82e9d3e085c83734d37b+}
Depends: libgnutlsxx28 (= [-3.5.8-5+deb9u3)-] {+3.5.8-5+deb9u4)+}
Version: [-3.5.8-5+deb9u3-] {+3.5.8-5+deb9u4+}

diff -Nru gnutls28-3.5.8/debian/changelog gnutls28-3.5.8/debian/changelog
--- gnutls28-3.5.8/debian/changelog 2017-07-23 14:28:37.0 +0200
+++ gnutls28-3.5.8/debian/changelog 2018-10-06 14:06:18.0 +0200
@@ -1,3 +1,14 @@
+gnutls28 (3.5.8-5+deb9u4) stretch; urgency=medium
+
+  * Pull fixes for CVE-2018-10844 and CVE-2018-10845 from gnutls 3.5.19
++ 39_01-dummy_wait-correctly-account-the-length-field-in-SHA.patch
++ 39_02-dummy_wait-always-hash-the-same-amount-of-blocks-tha.patch
++ 39_03-cbc_mac_verify-require-minimum-padding-under-SSL3.0.patch
++ 39_04-hmac-sha384-and-sha256-ciphersuites-were-removed-fro.patch
++ 39_05-tests-pkcs12_encode-fix-test-for-SHA512.patch
+
+ -- Andreas Metzler   Sat, 06 Oct 2018 14:06:18 +0200
+
 gnutls28 (3.5.8-5+deb9u3) stretch; urgency=medium
 
   * 38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch
diff -Nru 
gnutls28-3.5.8/debian/patches/39_01-dummy_wait-correctly-account-the-length-field-in-SHA.patch
 
gnutls28-3.5.8/debian/patches/39_01-dummy_wait-correctly-account-the-length-field-in-SHA.patch
--- 
gnutls28-3.5.8/debian/patches/39_01-dummy_wait-correctly-account-the-length-field-in-SHA.patch
  1970-01-01 01:00:00.0 +0100
+++ 
gnutls28-3.5.8/debian/patches/39_01-dummy_wait-correctly-account-the-length-field-in-SHA.patch
  2018-10-06 13:53:23.0 +0200
@@ -0,0 +1,92 @@
+From e14d85eb8b1987d86f7b1d101a0e7795675d20d4 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos 
+Date: Tue, 12 Jun 2018 14:22:52 +0200
+Subject: [PATCH 1/5] dummy_wait: correctly account the length field in SHA384
+ HMAC
+
+The existing lucky13 attack count-measures did not work correctly for
+SHA384 HMAC.
+
+The overall 

Bug#901551: unblock: libgcrypt20/1.8.3-1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package libgcrypt20, this is a upstream bugfix/security
release of the stable branch, fixing CVE-2018-0495.
https://www.nccgroup.trust/us/our-research/technical-advisory-return-of-the-hidden-number-problem/

unblock libgcrypt20/1.8.3-1

thanks, cu Andreas


1.8.2-2_to_1.8.3-1.diff.xz
Description: application/xz


signature.asc
Description: PGP signature

Bug#892802: transition: efl

[Andreas Metzler]

On 2018-04-08 Andreas Metzler <ametz...@bebt.de> wrote:
> On 2018-04-08 Emilio Pozuelo Monfort <po...@debian.org> wrote:
>> On 08/04/18 07:45, Andreas Metzler wrote:
>[...] 
>>> it looks like the transition needs some brute force/hint. Both efl and
>>> e17 are valid candidates, but do not propagate. Good somebody please
>>> take a look?

>> efl needs to be decrufted. For that, elementary should be removed
>> (AFAIK it's replaced by the new efl) so an RM bug against
>> ftp.debian.org is needed.

> Submitted as #895221

Status update:

+ elementary has been removed.
+ *Somehow* related exactimage/1.0.1-2 has dropped edisplay and
  therefore any efl-related dependency. It is only 2/5 days old. However
  this should not delay efl anymore since edisplay/testing's
  dependencies can still be fullfilled by efl/sid.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#892802: transition: efl

[Andreas Metzler]

On 2018-04-08 Emilio Pozuelo Monfort <po...@debian.org> wrote:
> On 08/04/18 07:45, Andreas Metzler wrote:
[...] 
>> it looks like the transition needs some brute force/hint. Both efl and
>> e17 are valid candidates, but do not propagate. Good somebody please
>> take a look?

> efl needs to be decrufted. For that, elementary should be removed
> (AFAIK it's replaced by the new efl) so an RM bug against
> ftp.debian.org is needed.

Submitted as #895221

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#892802: transition: efl

[Andreas Metzler]

On 2018-03-13 Ross Vandegrift  wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: transition

> Hello,

> I'd like to request a transition for efl from experimental -> unstable.  This
> release takes over a few other source packages.  It also reverses some
> Debian-local ABI & soname deviations from the upstream releases.
[...]

Hello,

it looks like the transition needs some brute force/hint. Both efl and
e17 are valid candidates, but do not propagate. Good somebody please
take a look?

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#892802: transition: efl

[Andreas Metzler]

On 2018-03-13 Emilio Pozuelo Monfort  wrote:
> Control: tags -1 confirmed

> On 13/03/18 08:15, Ross Vandegrift wrote:
>> Package: release.debian.org
>> Severity: normal
>> User: release.debian@packages.debian.org
>> Usertags: transition

>> Hello,

>> I'd like to request a transition for efl from experimental -> unstable.  This
>> release takes over a few other source packages.  It also reverses some
>> Debian-local ABI & soname deviations from the upstream releases.
[...]
> It looks good. Please go ahead and upload the affected packages to unstable,
> then I'll schedule the binNMUs for the "bad" packages listed in

> https://release.debian.org/transitions/html/auto-efl.html

Hello Emilio,

efl has been built on all archs it is going build[1], could you please
trigger the first run of binnmus?

FWIW the auto-efl tracker does not seem to work at all, I guess because
"good" and "bad" are not disjunct.

TIA, cu Andreas

[1] s390x is lacking luajit

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Re: mariadb-10.1 fails to propagate to testing

[Andreas Metzler]

On 2017-11-26 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> On Sun, 2017-11-26 at 16:22 +0100, Andreas Metzler wrote:
>> mariadb-10.1 1:10.1.29-6 seems to be stuck in sid. It does not
>> propagate to testing although
>> https://qa.debian.org/excuses.php?package=mariadb-10.1 lists it as
>> valid candidate.

>> Could you please check the cause?

> There's a bit of backstory, but effectively: the mariadb-test binary
> package in testing is built from the mariadb-10.1 source package, and
> has strictly versioned dependencies on other binaries built from that
> source package. In unstable, the mariadb-10.2 source package builds the
> binary package instead, but FTBFS on several architectures so is not a
> candidate. The net result is that when britney tries to migrate mariadb
> -10.1 1:10.1.29-6, the mariadb-test binary package in testing becomes
> uninstallable, and the migration attempt is aborted.

Hello,

So essentially if mariadb-10.1 migrated a mariadb-10.1 with mariadb-test
would be replaced by a mariadb-10.1 without mariadb-test?

Given that afaict mariadb-test has no reverse dependencies could you
consider forcing the migration?

I am asking since mariadb blocks exim, which urgently needs an update.

Thanks, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

mariadb-10.1 fails to propagate to testing

[Andreas Metzler]

Hello,

mariadb-10.1 1:10.1.29-6 seems to be stuck in sid. It does not propagate
to testing although
https://qa.debian.org/excuses.php?package=mariadb-10.1 lists it as valid
candidate.

Could you please check the cause?

Thanks, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#873855: unblock: libgcrypt20/1.7.9-1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package libgcrypt20, it includes the the fix for
CVE-2017-0379 https://www.debian.org/security/2017/dsa-3959

TIA, cu Andreas

unblock libgcrypt20/1.7.9-1

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#869434: stretch-pu: package gnutls28/3.5.8-5+deb9u3

[Andreas Metzler]

On 2017-08-08 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> On Sun, 2017-07-23 at 15:28 +0200, Andreas Metzler wrote:
>> gnutls upstream has pointed out that it would make sense to pull
>> two patches from 3.5.14. These improve interoperability by avoiding
>> incorrect OCSP verification errors. These errors could become quite
>> common with growing popularity of ecdsa signatures.

> Please go ahead.

Thanks, uploaded.

cu Andreas

Bug#869434: stretch-pu: package gnutls28/3.5.8-5+deb9u3

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

Hello,

gnutls upstream has pointed out that it would make sense to pull
two patches from 3.5.14. These improve interoperability by avoiding
incorrect OCSP verification errors. These errors could become quite
common with growing popularity of ecdsa signatures.

thanks, cu Andreas
diff -Nru gnutls28-3.5.8/debian/changelog gnutls28-3.5.8/debian/changelog
--- gnutls28-3.5.8/debian/changelog	2017-07-08 10:29:05.0 +0200
+++ gnutls28-3.5.8/debian/changelog	2017-07-23 14:28:37.0 +0200
@@ -1,3 +1,14 @@
+gnutls28 (3.5.8-5+deb9u3) stretch; urgency=medium
+
+  * 38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch
+38_02-OCSP-find_signercert-improved-DER-length-calculation.patch from
+gnutls 3.5.14: Fix OCSP verification errors, especially with ecdsa
+signatures.
+https://gitlab.com/gnutls/gnutls/issues/223
+Thanks to Nikos Mavrogiannopoulos for the suggestion.
+
+ -- Andreas Metzler <ametz...@debian.org>  Sun, 23 Jul 2017 14:28:37 +0200
+
 gnutls28 (3.5.8-5+deb9u2) stretch; urgency=medium
 
   * 37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch from
diff -Nru gnutls28-3.5.8/debian/patches/38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch gnutls28-3.5.8/debian/patches/38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch
--- gnutls28-3.5.8/debian/patches/38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.5.8/debian/patches/38_01-OCSP-check-the-subject-public-key-identifier-field-t.patch	2017-07-23 13:49:16.0 +0200
@@ -0,0 +1,56 @@
+From 4115dda443f38119ad46262f7f4adc78cfa1bf83 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Fri, 30 Jun 2017 10:04:01 +0200
+Subject: [PATCH 1/2] OCSP: check the subject public key identifier field to
+ figure issuer
+
+Normally when attempting to match the 'Responder Key ID' in an OCSP response
+against the issuer certificate we check (according to RFC6960) against the
+hash of the SPKI field. However, in few certificates (see commit:
+"added ECDSA OCSP response verification"), that may not be the case. In that
+certificate, that value matches the Subject Public Key identifier field
+but not the hash.
+
+To account for these certificates, we enhance the matching to also consider
+the Subject Public Key identifier field.
+
+Relates: #223
+
+Signed-off-by: Nikos Mavrogiannopoulos <n...@redhat.com>
+---
+ lib/x509/ocsp.c | 17 -
+ 1 file changed, 16 insertions(+), 1 deletion(-)
+
+diff --git a/lib/x509/ocsp.c b/lib/x509/ocsp.c
+index dcdf435d2..68e721eaa 100644
+--- a/lib/x509/ocsp.c
 b/lib/x509/ocsp.c
+@@ -1923,9 +1923,24 @@ static gnutls_x509_crt_t find_signercert(gnutls_ocsp_resp_t resp)
+ 
+ 	for (i = 0; i < ncerts; i++) {
+ 		if (keyid.data != NULL) {
+-			uint8_t digest[20];
++			uint8_t digest[128]; /* to support longer key IDs */
+ 			gnutls_datum_t spki;
++			size_t digest_size = sizeof(digest);
+ 
++			_gnutls_debug_log("checking key ID against SPK identifier\n");
++
++			/* check subject key identifier as well, some certificates
++			 * match that, but not the hash */
++			rc = gnutls_x509_crt_get_subject_key_id(certs[i], digest, _size, NULL);
++			if (rc >= 0 && digest_size == keyid.size &&
++			memcmp(keyid.data, digest, digest_size) == 0) {
++signercert = certs[i];
++goto quit;
++			}
++
++			_gnutls_debug_log("checking key ID against SPKI hash\n");
++
++			/* continue with checking the hash */
+ 			rc = _gnutls_x509_get_raw_field2(certs[i]->cert, [i]->der,
+ 	  "tbsCertificate.subjectPublicKeyInfo.subjectPublicKey",
+ 	  );
+-- 
+2.13.2
+
diff -Nru gnutls28-3.5.8/debian/patches/38_02-OCSP-find_signercert-improved-DER-length-calculation.patch gnutls28-3.5.8/debian/patches/38_02-OCSP-find_signercert-improved-DER-length-calculation.patch
--- gnutls28-3.5.8/debian/patches/38_02-OCSP-find_signercert-improved-DER-length-calculation.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.5.8/debian/patches/38_02-OCSP-find_signercert-improved-DER-length-calculation.patch	2017-07-23 13:49:16.0 +0200
@@ -0,0 +1,77 @@
+From 3c36d980d447251b34677c21bd4a141829c045f6 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@gnutls.org>
+Date: Sat, 1 Jul 2017 10:50:57 +0200
+Subject: [PATCH 2/2] OCSP: find_signercert: improved DER length calculation
+
+Previously we were assuming a fixed amount of length bytes which
+is not correct for all possible lengths. Use libtasn1 to decode
+the length field.
+
+Resolves: #223
+
+Signed-off-by: Nikos Mavrogiannopoulos <n...@gnutls.org>
+---
+ lib/x509/ocsp.c | 30 --
+ 1 file changed, 24 insertions(+), 6 deletions(-)
+
+diff --git a/lib/x509/ocsp.c b/lib/x509/ocsp.c
+in

Bug#867659: stretch-pu: package gnutls28/3.5.8-5+deb9u2

[Andreas Metzler]

On 2017-07-15 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> Control: tags -1 + confirmed

> On Sat, 2017-07-08 at 10:52 +0200, Andreas Metzler wrote:
>> I would like to fix #867581 in stable by pulling the patch from 3.5.13.
>> The issue is about broken AES-GCM in-place encryption and decryption on
>> aarch64.

> Please go ahead, bearing in mind that the window for 9.1 closes this
> weekend.

Thank you, uploaded.

cu Andreas

Bug#867665: transition: wmaker

[Andreas Metzler]

On 2017-07-09 Jonathan Wiltshire <j...@debian.org> wrote:
> Control: tag -1 confirmed

> On Sat, Jul 08, 2017 at 01:22:26PM +0200, Andreas Metzler wrote:
>> wmaker 0.95.8 features a soname bump of libwraster. There are only 3
>> other packages involved (wdm, wmforecast and wmweather+) and all of them
>> build fine against the libwmaker-dev in experimental.

> Please go ahead.

Thanks, I have just uploaded to sid.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#867665: transition: wmaker

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

wmaker 0.95.8 features a soname bump of libwraster. There are only 3
other packages involved (wdm, wmforecast and wmweather+) and all of them
build fine against the libwmaker-dev in experimental.

The auto-wmaker tracker looks fine, here is the Ben file generated by
reportbug:
title = "wmaker";
is_affected = .depends ~ "libwraster5" | .depends ~ "libwraster6";
is_good = .depends ~ "libwraster6";
is_bad = .depends ~ "libwraster5";

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'


signature.asc
Description: PGP signature

Bug#867659: stretch-pu: package gnutls28/3.5.8-5+deb9u2

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

Hello,

I would like to fix #867581 in stable by pulling the patch from 3.5.13.
The issue is about broken AES-GCM in-place encryption and decryption on
aarch64.

cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.5.8/debian/changelog gnutls28-3.5.8/debian/changelog
--- gnutls28-3.5.8/debian/changelog	2017-06-16 07:47:11.0 +0200
+++ gnutls28-3.5.8/debian/changelog	2017-07-08 10:29:05.0 +0200
@@ -1,3 +1,11 @@
+gnutls28 (3.5.8-5+deb9u2) stretch; urgency=medium
+
+  * 37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch from
+upstream 3.5.x branch: Fix breakage if AES-GCM in-place encryption and
+decryption on aarch64. Closes: #867581
+
+ -- Andreas Metzler <ametz...@debian.org>  Sat, 08 Jul 2017 10:29:05 +0200
+
 gnutls28 (3.5.8-5+deb9u1) stretch-security; urgency=high
 
   * 36_CVE-2017-7507_*.patch: Pulled from 3.5.13, fix crash upon receiving
diff -Nru gnutls28-3.5.8/debian/patches/37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch gnutls28-3.5.8/debian/patches/37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch
--- gnutls28-3.5.8/debian/patches/37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.5.8/debian/patches/37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch	2017-07-07 19:43:55.0 +0200
@@ -0,0 +1,57 @@
+From 864e8d4e3ba87f53df7bdef695661415ed60a018 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Mon, 22 May 2017 14:41:56 +0200
+Subject: [PATCH] aarch64: fix AES-GCM in-place encryption and decryption
+
+Resolves #204
+
+Signed-off-by: Nikos Mavrogiannopoulos <n...@redhat.com>
+---
+ lib/accelerated/aarch64/aes-gcm-aarch64.c | 24 
+ 1 file changed, 24 insertions(+)
+
+diff --git a/lib/accelerated/aarch64/aes-gcm-aarch64.c b/lib/accelerated/aarch64/aes-gcm-aarch64.c
+index c571d0294..8d2bc1dce 100644
+--- a/lib/accelerated/aarch64/aes-gcm-aarch64.c
 b/lib/accelerated/aarch64/aes-gcm-aarch64.c
+@@ -153,6 +153,27 @@ gcm_ghash(struct aes_gcm_ctx *ctx, const uint8_t * src, size_t src_size)
+ }
+ 
+ static void
++ctr32_encrypt_blocks_inplace(const unsigned char *in, unsigned char *out,
++			 size_t blocks, const AES_KEY *key,
++			 const unsigned char ivec[16])
++{
++	unsigned i;
++	uint8_t ctr[16];
++	uint8_t tmp[16];
++
++	memcpy(ctr, ivec, 16);
++
++	for (i=0;i<blocks;i++) {
++		aes_v8_encrypt(ctr, tmp, key);
++		memxor3(out, tmp, in, 16);
++
++		out += 16;
++		in += 16;
++		INCREMENT(16, ctr);
++	}
++}
++
++static void
+ ctr32_encrypt_blocks(const unsigned char *in, unsigned char *out,
+ 		 size_t blocks, const AES_KEY *key,
+ 		 const unsigned char ivec[16])
+@@ -160,6 +181,9 @@ ctr32_encrypt_blocks(const unsigned char *in, unsigned char *out,
+ 	unsigned i;
+ 	uint8_t ctr[16];
+ 
++	if (in == out)
++		return ctr32_encrypt_blocks_inplace(in, out, blocks, key, ivec);
++
+ 	memcpy(ctr, ivec, 16);
+ 
+ 	for (i=0;i<blocks;i++) {
+-- 
+2.13.2
+
diff -Nru gnutls28-3.5.8/debian/patches/series gnutls28-3.5.8/debian/patches/series
--- gnutls28-3.5.8/debian/patches/series	2017-06-16 07:47:04.0 +0200
+++ gnutls28-3.5.8/debian/patches/series	2017-07-07 19:43:58.0 +0200
@@ -12,3 +12,4 @@
 36_CVE-2017-7507_1-ext-status_request-ensure-response-IDs-are-properly-.patch
 36_CVE-2017-7507_2-ext-status_request-Removed-the-parsing-of-responder-.patch
 36_CVE-2017-7507_3-gnutls_ocsp_status_request_enable_client-documented-.patch
+37_aarch64-fix-AES-GCM-in-place-encryption-and-decrypti.patch


signature.asc
Description: PGP signature

Bug#865763: jessie-pu: package gnutls28/3.3.8-6+deb8u7

[Andreas Metzler]

On 2017-06-27 Cyril Brulebois <k...@debian.org> wrote:
> Andreas Metzler <ametz...@bebt.de> (2017-06-24):
>> would like to fix the following issue in gnutls28/jessie (It was fixed
>> in 3.5.3 and therefore does not apply to stretch/buster/sid).
 
>> Quoting #865297:
>> 
>> If the application closes open files during startup (e.g., a daemon),
>> it may close the file that gnutls has open for /dev/urandom. The
>> recommended way to handle this situation is to call
>> gnutls_global_init() again. This will check if the fd for /dev/urandom
>> is still valid and re-open it if not.
>> 
>> Unfortunately, the way that the /dev/urandom fd is checked is not
>> reliable. It only checks the mode, which might be the same if the
>> application reused the fd for another character device with the same
>> permissions (e.g., /dev/null).
>> 

> The patch looks good to me, but I'd like to get a clarification: is the
> fix in 3.5.3 based on the same patch, or was a different route taken?

Yes, the same route was taken. The patch on the gnutls_3_3_x branch
5006914fda50f25807451a03616cdf2e7be0268f was picked and unfuzzed from
408cfd7a3afba0c5a2310c5cbcee581f57d9248c on gnutls_3_5_x

> I'd like to avoid letting something go through (o-)p-u that hasn't seen
> much testing elsewhere.

Understandable. ;-)

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#865763: jessie-pu: package gnutls28/3.3.8-6+deb8u7

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu

Hello,

would like to fix the following issue in gnutls28/jessie (It was fixed
in 3.5.3 and therefore does not apply to stretch/buster/sid).

Quoting #865297:

If the application closes open files during startup (e.g., a daemon),
it may close the file that gnutls has open for /dev/urandom. The
recommended way to handle this situation is to call
gnutls_global_init() again. This will check if the fd for /dev/urandom
is still valid and re-open it if not.

Unfortunately, the way that the /dev/urandom fd is checked is not
reliable. It only checks the mode, which might be the same if the
application reused the fd for another character device with the same
permissions (e.g., /dev/null).


Thanks for considering, cu Andreas
diff -Nru gnutls28-3.3.8/debian/changelog gnutls28-3.3.8/debian/changelog
--- gnutls28-3.3.8/debian/changelog	2017-06-16 07:39:56.0 +0200
+++ gnutls28-3.3.8/debian/changelog	2017-06-24 17:50:29.0 +0200
@@ -1,3 +1,13 @@
+gnutls28 (3.3.8-6+deb8u7) jessie; urgency=medium
+
+  * 57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch from
+upstream gnutls_3_3_x branch: Improve check for /dev/urandom uniqueness.
+Ensure that when gnutls_global_init() is called for a second time that
+/dev/urandom is re-opened when the inode or device ID has changed.
+Closes: #865297
+
+ -- Andreas Metzler <ametz...@debian.org>  Sat, 24 Jun 2017 17:50:21 +0200
+
 gnutls28 (3.3.8-6+deb8u6) jessie-security; urgency=high
 
   * 56_CVE-2017-7507_1-ext-status_request-ensure-response-IDs-are-pro.patch
diff -Nru gnutls28-3.3.8/debian/patches/57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch gnutls28-3.3.8/debian/patches/57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch
--- gnutls28-3.3.8/debian/patches/57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.3.8/debian/patches/57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch	2017-06-24 17:49:20.0 +0200
@@ -0,0 +1,56 @@
+From 5006914fda50f25807451a03616cdf2e7be0268f Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Fri, 15 Jul 2016 14:58:07 +0200
+Subject: [PATCH] urandom: use st_ino and st_rdev to determine device
+ uniqueness
+
+---
+ lib/nettle/rnd-common.c | 11 +++
+ 1 file changed, 7 insertions(+), 4 deletions(-)
+
+diff --git a/lib/nettle/rnd-common.c b/lib/nettle/rnd-common.c
+index 47d2b0edd..33a71d351 100644
+--- a/lib/nettle/rnd-common.c
 b/lib/nettle/rnd-common.c
+@@ -137,7 +137,8 @@ void _rnd_system_entropy_deinit(void)
+ #include "egd.h"
+ 
+ static int _gnutls_urandom_fd = -1;
+-static mode_t _gnutls_urandom_fd_mode = 0;
++static ino_t _gnutls_urandom_fd_ino = 0;
++static dev_t _gnutls_urandom_fd_rdev = 0;
+ 
+ static int _rnd_get_system_entropy_urandom(void* _rnd, size_t size)
+ {
+@@ -202,7 +203,7 @@ int _rnd_system_entropy_check(void)
+ 	struct stat st;
+ 
+ 	ret = fstat(_gnutls_urandom_fd, );
+-	if (ret < 0 || st.st_mode != _gnutls_urandom_fd_mode) {
++	if (ret < 0 || st.st_ino != _gnutls_urandom_fd_ino || st.st_rdev != _gnutls_urandom_fd_rdev) {
+ 		return _rnd_system_entropy_init();
+ 	}
+ 	return 0;
+@@ -224,7 +225,8 @@ int _rnd_system_entropy_init(void)
+ 		fcntl(_gnutls_urandom_fd, F_SETFD, old | FD_CLOEXEC);
+ 
+ 	if (fstat(_gnutls_urandom_fd, ) >= 0) {
+-		_gnutls_urandom_fd_mode = st.st_mode;
++		_gnutls_urandom_fd_ino = st.st_ino;
++		_gnutls_urandom_fd_rdev = st.st_rdev;
+ 	}
+ 
+ 	_rnd_get_system_entropy = _rnd_get_system_entropy_urandom;
+@@ -240,7 +242,8 @@ fallback:
+ 	}
+ 
+ 	if (fstat(_gnutls_urandom_fd, ) >= 0) {
+-		_gnutls_urandom_fd_mode = st.st_mode;
++		_gnutls_urandom_fd_ino = st.st_ino;
++		_gnutls_urandom_fd_rdev = st.st_rdev;
+ 	}
+ 
+ 	_rnd_get_system_entropy = _rnd_get_system_entropy_egd;
+-- 
+2.11.0
+
diff -Nru gnutls28-3.3.8/debian/patches/series gnutls28-3.3.8/debian/patches/series
--- gnutls28-3.3.8/debian/patches/series	2017-06-15 16:13:12.0 +0200
+++ gnutls28-3.3.8/debian/patches/series	2017-06-24 17:50:33.0 +0200
@@ -36,3 +36,4 @@
 56_CVE-2017-7507_1-ext-status_request-ensure-response-IDs-are-pro.patch
 56_CVE-2017-7507_2-ext-status_request-Removed-the-parsing-of-resp.patch
 56_CVE-2017-7507_3-gnutls_ocsp_status_request_enable_client-docum.patch
+57_urandom-use-st_ino-and-st_rdev-to-determine-device-u.patch


signature.asc
Description: PGP signature

Bug#864968: transition: libunistring

[Andreas Metzler]

On 2017-06-18 Jörg Frings-Fürst  wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: transition

> Hello,

> I'd want to upload libunistring 0.9.7 to unstable. The ABI is bumped.
[...]

Hello,

Point of information: Afaict from the changelog the only
reason for the soname bump in 0.9.4 is this one:
-
diff -NurbBp xxx-libunistring-dev_0.9.6+really0.9.3-0.1/usr/include/unistr.h 
xxx-libunistring-dev_0.9.7-1/usr/include/unistr.h
--- xxx-libunistring-dev_0.9.6+really0.9.3-0.1/usr/include/unistr.h 
2016-05-27 12:54:58.0 +0200
+++ xxx-libunistring-dev_0.9.7-1/usr/include/unistr.h   2017-06-10 
12:18:01.0 +0200
@@ -562,12 +595,23 @@ extern uint32_t *
 
 /* Compare S1 and S2.  */
 /* Similar to strcmp(), wcscmp().  */
+#ifdef __sun
+/* Avoid a collision with the u8_strcmp() function in Solaris 11 libc.  */
+extern int
+   u8_strcmp_gnu (const uint8_t *s1, const uint8_t *s2)
+   _UC_ATTRIBUTE_PURE;
+# define u8_strcmp u8_strcmp_gnu
+#else
 extern int
-   u8_strcmp (const uint8_t *s1, const uint8_t *s2);
+   u8_strcmp (const uint8_t *s1, const uint8_t *s2)
+   _UC_ATTRIBUTE_PURE;
+#endif
-
i.e. it is specific to __sun. So afaict this transition should go
smoothly without breakage of programs linking against both version of
libunistring although libunistring does not use versioned symbols.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

unblocking for stretch point release?

[Andreas Metzler]

Hello,

I first understood the latest mail to -announce ("Planned release of
stretch") to mean that propagation from sid to stretch is not possible
anymore (except for critical fixes).

However now that I am in a position of wanting to get something into the
1st point release I am wondering whether that is true (and I need to
prepare an additional upload for proposed updated) or whether
propagation from sid to stretch r1 is possible.

cu Andreas
PS: (Just for completeness sake: This mail was triggered by Gnutls
CVE-2017-7507.)
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#864083: unblock: libgcrypt20/1.7.6-2

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package libgcrypt20, the upload features the following
changes:
* Refresh debian/upstream/signing-key.asc, key-expiry-dates bumped.
* Pull two fixes from gcrypt 1.7.7 bugfix release:
  + 30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch
Fix possible timing attack on EdDSA session key.
  + 30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch
Fix long standing bug in secure memory implementation which could lead
to a segv on free.

unblock libgcrypt20/1.7.6-2

Thanks, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru libgcrypt20-1.7.6/debian/changelog libgcrypt20-1.7.6/debian/changelog
--- libgcrypt20-1.7.6/debian/changelog	2017-01-26 11:58:32.0 +0100
+++ libgcrypt20-1.7.6/debian/changelog	2017-06-03 10:58:36.0 +0200
@@ -1,3 +1,15 @@
+libgcrypt20 (1.7.6-2) unstable; urgency=high
+
+  * Refresh debian/upstream/signing-key.asc, key-expiry-dates bumped.
+  * Pull two fixes from gcrypt 1.7.7 bugfix release:
++ 30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch
+  Fix possible timing attack on EdDSA session key.
++ 30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch
+  Fix long standing bug in secure memory implementation which could lead
+  to a segv on free.
+
+ -- Andreas Metzler <ametz...@debian.org>  Sat, 03 Jun 2017 10:58:36 +0200
+
 libgcrypt20 (1.7.6-1) unstable; urgency=medium
 
   * New upstream version, includes
diff -Nru libgcrypt20-1.7.6/debian/patches/30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch libgcrypt20-1.7.6/debian/patches/30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch
--- libgcrypt20-1.7.6/debian/patches/30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch	1970-01-01 01:00:00.0 +0100
+++ libgcrypt20-1.7.6/debian/patches/30_gcry177_01-ecc-Store-EdDSA-session-key-in-secure-memory.patch	2017-06-03 10:53:37.0 +0200
@@ -0,0 +1,35 @@
+From f9494b3f258e01b6af8bd3941ce436bcc00afc56 Mon Sep 17 00:00:00 2001
+From: Jo Van Bulck <jo.vanbu...@cs.kuleuven.be>
+Date: Thu, 19 Jan 2017 17:00:15 +0100
+Subject: [PATCH 1/2] ecc: Store EdDSA session key in secure memory.
+
+* cipher/ecc-eddsa.c (_gcry_ecc_eddsa_sign): use mpi_snew to allocate
+session key.
+--
+
+An attacker who learns the EdDSA session key from side-channel
+observation during the signing process, can easily revover the long-
+term secret key. Storing the session key in secure memory ensures that
+constant time point operations are used in the MPI library.
+
+Signed-off-by: Jo Van Bulck <jo.vanbu...@cs.kuleuven.be>
+---
+ cipher/ecc-eddsa.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/cipher/ecc-eddsa.c b/cipher/ecc-eddsa.c
+index f91f8489..813e030d 100644
+--- a/cipher/ecc-eddsa.c
 b/cipher/ecc-eddsa.c
+@@ -603,7 +603,7 @@ _gcry_ecc_eddsa_sign (gcry_mpi_t input, ECC_secret_key *skey,
+   a = mpi_snew (0);
+   x = mpi_new (0);
+   y = mpi_new (0);
+-  r = mpi_new (0);
++  r = mpi_snew (0);
+   ctx = _gcry_mpi_ec_p_internal_new (skey->E.model, skey->E.dialect, 0,
+  skey->E.p, skey->E.a, skey->E.b);
+   b = (ctx->nbits+7)/8;
+-- 
+2.11.0
+
diff -Nru libgcrypt20-1.7.6/debian/patches/30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch libgcrypt20-1.7.6/debian/patches/30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch
--- libgcrypt20-1.7.6/debian/patches/30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch	1970-01-01 01:00:00.0 +0100
+++ libgcrypt20-1.7.6/debian/patches/30_gcry177_02-secmem-Fix-SEGV-and-stat-calculation.patch	2017-06-03 10:53:37.0 +0200
@@ -0,0 +1,69 @@
+From 91456759b887e153c4d4ce19538d478df260cab2 Mon Sep 17 00:00:00 2001
+From: NIIBE Yutaka <gni...@fsij.org>
+Date: Fri, 2 Jun 2017 10:34:42 +0900
+Subject: [PATCH 2/2] secmem: Fix SEGV and stat calculation.
+
+* src/secmem (init_pool): Care about the header size.
+(_gcry_secmem_malloc_internal): Likewise.
+(_gcry_secmem_malloc_internal): Use mb->size for stats.
+
+--
+
+GnuPG-bug-id: 3027
+Signed-off-by: NIIBE Yutaka <gni...@fsij.org>
+---
+ src/secmem.c | 10 +-
+ 1 file changed, 5 insertions(+), 5 deletions(-)
+
+diff --git a/src/secmem.c b/src/secmem.c
+index 46bbf82e..b2a9667d 100644
+--- a/src/secmem.c
 b/src/secmem.c
+@@ -454,7 +454,7 @@ init_pool (pooldesc_t *pool, size_t n)
+ 
+   /* Initialize first memory block.  */
+   mb = (memblock_t *) pool->mem;
+-  mb->size = pool->size;
++  mb->size = pool->size - BLOCK_HEAD_SIZE;
+   mb->flags = 0;
+ }
+ 
+@@ -610,7 +610,7 @@ _gcry_secmem_malloc_internal (size_t size, int xhint)
+   mb = mb_get_new (pool, (memblock_t *) pool->mem, size);
+   if (mb)
+ {
+-   

Bug#856872: jessie-pu: package gnutls28/3.3.8-6+deb8u5

[Andreas Metzler]

On 2017-04-27 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> On Mon, 2017-03-06 at 19:24 +0100, Andreas Metzler wrote:
[...]
>> upstream has now released 3.5.10/3.3.27 including these fixes and
>> another one on top:
>>  + 55_16_Enforce-the-max-packet-length-for-OpenPGP-subpackets.patch
>>Addressed large allocation in OpenPGP certificate parsing, that could
>>lead in out-of-memory condition. Issue found using oss-fuzz project, 
>> and
>>was fixed by Alex Gaynor:
>>https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=392
>>[GNUTLS-SA-2017-3C]
>> 
>> Updated diff for jessie attached.

> Please go ahead; thanks.

Thanks, uploaded with the new CVE number mentioned in changelog.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#856872: jessie-pu: package gnutls28/3.3.8-6+deb8u5

[Andreas Metzler]

On 2017-04-23 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> On Sun, 2017-03-05 at 19:08 +0100, Andreas Metzler wrote:
> > I would like fix a number of minor issues in GnuTLS.

> Apologies for the delay in getting back to you.

> Are all of the issues listed below already resolved in unstable?

>> Most of these (notably CVE-2017-533[4567]) are related to the PGP
>> support, security does not intend to issue a DSA:
>> 
>> + 55_00_pkcs12-fixed-the-calculation-of-p_size.patch
>>   Fixed issue in PKCS#12 password encoding, which truncated
>>   passwords over 32-characters. Reported by Mario Klebsch.
[1]
[...]

Hello Adam,

Yes, all of these are fixed in 3.5.8-5, which is also in testing.

cu Andreas

[1] Could not quickly find a commit on the 3.5 branch for this specific
issue but have verified that it does not apply to 3.5.8. I guess it
never applied to 3.5.x.
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#857460: unblock: exim4/4.89-1

[Andreas Metzler]

On 2017-03-11 Andreas Metzler <ametz...@bebt.de> wrote:
[...]
> unblock exim4/4.89-1
> unblock eximdoc4/4.89-1

Ping?

Bug#857292: unblock: hugin/2016.2.0+dfsg-1

[Andreas Metzler]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package hugin, this fixes two issues:
* #822062, #855505: When a custom temp directory was sent in the
  preferences the hugin assistant failed when aligning pictures.
* Bumped version number. The tarball is *binary* *identical* with rc2.

thanks, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru hugin-2016.2.0~rc2+dfsg/debian/changelog hugin-2016.2.0+dfsg/debian/changelog
--- hugin-2016.2.0~rc2+dfsg/debian/changelog	2016-09-11 13:40:11.0 +0200
+++ hugin-2016.2.0+dfsg/debian/changelog	2017-02-26 08:23:23.0 +0100
@@ -1,3 +1,14 @@
+hugin (2016.2.0+dfsg-1) unstable; urgency=medium
+
+  * rc2 released as 2016.2.0.
+  * Fix assissant align error with custom temporary directory
+50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff
+  error.
+51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff
+Closes: #822062, #855505
+
+ -- Andreas Metzler <ametz...@debian.org>  Sun, 26 Feb 2017 08:23:23 +0100
+
 hugin (2016.2.0~rc2+dfsg-2) unstable; urgency=medium
 
   * Upload to unstable.
diff -Nru hugin-2016.2.0~rc2+dfsg/debian/patches/50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff hugin-2016.2.0+dfsg/debian/patches/50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff
--- hugin-2016.2.0~rc2+dfsg/debian/patches/50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff	1970-01-01 01:00:00.0 +0100
+++ hugin-2016.2.0+dfsg/debian/patches/50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff	2016-12-18 15:05:14.0 +0100
@@ -0,0 +1,20 @@
+# HG changeset patch
+# User tmodes
+# Date 1474121697 -7200
+#  Sat Sep 17 16:14:57 2016 +0200
+# Node ID 67c64f0ca1c45c991efa7b11e457cc5bc2c945c0
+# Parent  91b78c0c991f60744d71c9a4bd434c46cba4b444
+Fixes a type of format string
+
+diff -r 91b78c0c991f -r 67c64f0ca1c4 src/hugin1/hugin/MainFrame.cpp
+--- a/src/hugin1/hugin/MainFrame.cpp	Sat Sep 17 16:14:24 2016 +0200
 b/src/hugin1/hugin/MainFrame.cpp	Sat Sep 17 16:14:57 2016 +0200
+@@ -2162,7 +2162,7 @@
+ unsigned i2 = *(comps[1].begin());
+ ShowCtrlPointEditor( i1, i2);
+ // display message box with 
+-wxMessageBox(wxString::Format(_("Warning %d unconnected image groups found:"), comps.size()) + Components2Str(comps) + wxT("\n")
++wxMessageBox(wxString::Format(_("Warning %d unconnected image groups found:"), static_cast(comps.size())) + Components2Str(comps) + wxT("\n")
+ + _("Please create control points between unconnected images using the Control Points tab in the panorama editor.\n\nAfter adding the points, press the \"Align\" button again"),_("Error"), wxOK , mainWin);
+ return;
+ };
diff -Nru hugin-2016.2.0~rc2+dfsg/debian/patches/51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff hugin-2016.2.0+dfsg/debian/patches/51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff
--- hugin-2016.2.0~rc2+dfsg/debian/patches/51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff	1970-01-01 01:00:00.0 +0100
+++ hugin-2016.2.0+dfsg/debian/patches/51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff	2017-02-26 08:05:34.0 +0100
@@ -0,0 +1,20 @@
+# HG changeset patch
+# User tmodes
+# Date 1488046377 -3600
+#  Sat Feb 25 19:12:57 2017 +0100
+# Node ID 401823447b218b6d4778664fb6496deea6b03936
+# Parent  0fec458e26560958d657d14fd9a7a99b7f5c6c10
+Fixes running assistant with user defined temp directory [1666030]
+
+--- a/src/hugin1/base_wx/MyExternalCmdExecDialog.cpp
 b/src/hugin1/base_wx/MyExternalCmdExecDialog.cpp
+@@ -255,6 +255,9 @@ int MyExecPanel::ExecQueue(HuginQueue::C
+ #if wxCHECK_VERSION(3,0,0)
+ wxConfigBase* config = wxConfigBase::Get();
+ const long threads = config->Read(wxT("/output/NumberOfThreads"), 0l);
++// read all current environment variables
++wxGetEnvMap(_executeEnv.env);
++// now modify some variables before passing them to wxExecute
+ if (threads > 0)
+ {
+ wxString s;
diff -Nru hugin-2016.2.0~rc2+dfsg/debian/patches/series hugin-2016.2.0+dfsg/debian/patches/series
--- hugin-2016.2.0~rc2+dfsg/debian/patches/series	2016-07-16 07:03:27.0 +0200
+++ hugin-2016.2.0+dfsg/debian/patches/series	2017-02-26 08:00:14.0 +0100
@@ -1 +1,3 @@
 43_fallbackhelp.patch
+50_67c64f0ca1c4_Fixes_a_type_of_format_string.diff
+51_401823447b21_Fixes_running_assistant_with_user_defined_temp_directory.diff


signature.asc
Description: PGP signature

Bug#856872: jessie-pu: package gnutls28/3.3.8-6+deb8u5

[Andreas Metzler]

On 2017-03-05 Andreas Metzler <ametz...@bebt.de> wrote:
> Package: release.debian.org
> Severity: normal
> Tags: jessie
> User: release.debian@packages.debian.org
> Usertags: pu

> Hello,

> I would like fix a number of minor issues in GnuTLS.

> Most of these (notably CVE-2017-533[4567]) are related to the PGP
> support, security does not intend to issue a DSA:
[...]

Hello,

upstream has now released 3.5.10/3.3.27 including these fixes and
another one on top:
 + 55_16_Enforce-the-max-packet-length-for-OpenPGP-subpackets.patch
   Addressed large allocation in OpenPGP certificate parsing, that could
   lead in out-of-memory condition. Issue found using oss-fuzz project, and
   was fixed by Alex Gaynor:
   https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=392
   [GNUTLS-SA-2017-3C]

Updated diff for jessie attached.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.3.8/debian/changelog gnutls28-3.3.8/debian/changelog
--- gnutls28-3.3.8/debian/changelog	2016-11-01 10:07:52.0 +0100
+++ gnutls28-3.3.8/debian/changelog	2017-03-06 19:13:23.0 +0100
@@ -1,3 +1,62 @@
+gnutls28 (3.3.8-6+deb8u5) jessie; urgency=medium
+
+  * Pull multiple fixes from gnutls_3_3_x branch:
++ 55_00_pkcs12-fixed-the-calculation-of-p_size.patch
+  Fixed issue in PKCS#12 password encoding, which truncated
+  passwords over 32-characters. Reported by Mario Klebsch.
++ 55_01_gnutls_x509_ext_import_proxy-fix-issue-reading-the-p.patch
+  Fix double free in certificate information printing. If the PKIX
+  extension proxy was set with a policy language set but no policy
+  specified, that could lead to a double free. [GNUTLS-SA-2017-1]
+  CVE-2017-5334
++ 55_02_auth-rsa-eliminated-memory-leak-on-pkcs-1-formatting.patch
+  Addressed memory leak in server side error path (issue found using
+  oss-fuzz project)
++ 55_03_opencdk-Fixes-to-prevent-undefined-behavior-found-wi.patch
+  55_04_Do-not-infinite-loop-if-an-EOF-occurs-while-skipping.patch
+  55_05_Attempt-to-fix-a-leak-in-OpenPGP-cert-parsing.patch
+  55_06_Corrected-a-leak-in-OpenPGP-sub-packet-parsing.patch
+  55_07_opencdk-read_attribute-added-more-precise-checks-whe.patch
+  55_08_opencdk-cdk_pk_get_keyid-fix-stack-overflow.patch
+  55_09_opencdk-added-error-checking-in-the-stream-reading-f.patch
+  55_10_opencdk-improved-error-code-checking-in-the-stream-r.patch
+  55_11_opencdk-read-packet.c-corrected-typo-in-type-cast.patch
+  Addressed memory leaks and an infinite loop in OpenPGP certificate
+  parsing. Fixes by Alex Gaynor. (issues found using oss-fuzz project)
+  Addressed invalid memory accesses in OpenPGP certificate parsing.
+  (issues found using oss-fuzz project) [GNUTLS-SA-2017-2]
+  CVE-2017-5335 / CVE-2017-5336 / CVE-2017-5337
++ 55_12_gnutls_pkcs11_obj_list_import_url2-Always-return-an-.patch
+  When returning success, but no elements,
+  gnutls_pkcs11_obj_list_import_url4, could have returned zero number of
+  elements with a pointer that was uninitialized. Ensure that an
+  initialized (i.e., null in that case), pointer is always returned.
++ 55_13_cdk_pkt_read-enforce-packet-limits.patch Addressed integer
+  overflow resulting to invalid memory write in OpenPGP certificate
+  parsing.  Issue found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=420
+  [GNUTLS-SA-2017-3A]
++ 55_14_opencdk-read_attribute-account-buffer-size.patch Addressed read
+  of 1 byte past the end of buffer in OpenPGP certificate parsing. Issue
+  found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=391
+  (This patch is from gnutls_3_5_x branch.)
++ 55_15_opencdk-do-not-parse-any-secret-keys-in-packet-when-.patch
+  Addressed crashes in OpenPGP certificate parsing, related to private key
+  parser. No longer allow OpenPGP certificates (public keys) to contain
+  private key sub-packets. Issue found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=354
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=360
+  [GNUTLS-SA-2017-3B]
++ 55_16_Enforce-the-max-packet-length-for-OpenPGP-subpackets.patch
+  Addressed large allocation in OpenPGP certificate parsing, that could
+  lead in out-of-memory condition. Issue found using oss-fuzz project, and
+  was fixed by Alex Gaynor:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=392
+  [GNUTLS-SA-2017-3C]
+
+ -- Andreas Metzler <ametz...@debian.org>  Mon, 06 Mar 2017 19:13:20 +0100
+
 gnutls28 (3.3.8-6+deb8u4) jessie; urgency=medium
 
   [ Salvatore Bonaccorso ]
diff -Nru gnutls28-3.3.8/debian/patches/55_00_pkcs12-fixed-the-calculatio

Bug#856872: jessie-pu: package gnutls28/3.3.8-6+deb8u5

[Andreas Metzler]

 project)
+  Addressed invalid memory accesses in OpenPGP certificate parsing.
+  (issues found using oss-fuzz project) [GNUTLS-SA-2017-2]
+  CVE-2017-5335 / CVE-2017-5336 / CVE-2017-5337
++ 55_12_gnutls_pkcs11_obj_list_import_url2-Always-return-an-.patch
+  When returning success, but no elements,
+  gnutls_pkcs11_obj_list_import_url4, could have returned zero number of
+  elements with a pointer that was uninitialized. Ensure that an
+  initialized (i.e., null in that case), pointer is always returned.
++ 55_13_cdk_pkt_read-enforce-packet-limits.patch Addressed integer
+  overflow resulting to invalid memory write in OpenPGP certificate
+  parsing.  Issue found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=420
+  [GNUTLS-SA-2017-3A]
++ 55_14_opencdk-read_attribute-account-buffer-size.patch Addressed read
+  of 1 byte past the end of buffer in OpenPGP certificate parsing. Issue
+  found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=391
+  (This patch is from gnutls_3_5_x branch.)
++ 55_15_opencdk-do-not-parse-any-secret-keys-in-packet-when-.patch
+  Addressed crashes in OpenPGP certificate parsing, related to private key
+  parser. No longer allow OpenPGP certificates (public keys) to contain
+  private key sub-packets. Issue found using oss-fuzz project:
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=354
+  https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=360
+  [GNUTLS-SA-2017-3B]
+
+ -- Andreas Metzler <ametz...@debian.org>  Sun, 05 Mar 2017 18:18:03 +0100
+
 gnutls28 (3.3.8-6+deb8u4) jessie; urgency=medium
 
   [ Salvatore Bonaccorso ]
diff -Nru gnutls28-3.3.8/debian/patches/55_00_pkcs12-fixed-the-calculation-of-p_size.patch gnutls28-3.3.8/debian/patches/55_00_pkcs12-fixed-the-calculation-of-p_size.patch
--- gnutls28-3.3.8/debian/patches/55_00_pkcs12-fixed-the-calculation-of-p_size.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.3.8/debian/patches/55_00_pkcs12-fixed-the-calculation-of-p_size.patch	2017-03-05 08:29:46.0 +0100
@@ -0,0 +1,26 @@
+From 3979cbcb425b4088c822b0a75c78f5f1eef32291 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Mon, 28 Nov 2016 11:47:40 +0100
+Subject: [PATCH] pkcs12: fixed the calculation of p_size
+
+That affects passwords which exceed 32 characters.
+---
+ lib/x509/pkcs12_encr.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/lib/x509/pkcs12_encr.c b/lib/x509/pkcs12_encr.c
+index 85cd3f228..d8fd49f82 100644
+--- a/lib/x509/pkcs12_encr.c
 b/lib/x509/pkcs12_encr.c
+@@ -105,7 +105,7 @@ _gnutls_pkcs12_string_to_key(const mac_entry_st * me,
+ 	}
+ 
+ 	/* Store salt and password in BUF_I */
+-	p_size = ((pwlen / 64) * 64) + 64;
++	p_size = (((2*pwlen) / 64) * 64) + 64;
+ 
+ 	if (p_size > sizeof(buf_i) - 64)
+ 		return gnutls_assert_val(GNUTLS_E_INVALID_REQUEST);
+-- 
+2.11.0
+
diff -Nru gnutls28-3.3.8/debian/patches/55_01_gnutls_x509_ext_import_proxy-fix-issue-reading-the-p.patch gnutls28-3.3.8/debian/patches/55_01_gnutls_x509_ext_import_proxy-fix-issue-reading-the-p.patch
--- gnutls28-3.3.8/debian/patches/55_01_gnutls_x509_ext_import_proxy-fix-issue-reading-the-p.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.3.8/debian/patches/55_01_gnutls_x509_ext_import_proxy-fix-issue-reading-the-p.patch	2017-03-05 08:29:46.0 +0100
@@ -0,0 +1,76 @@
+From bbfd47d4bb6935b3eddae227deb9f340e2c1a69d Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Thu, 15 Dec 2016 15:02:18 +0100
+Subject: [PATCH] gnutls_x509_ext_import_proxy: fix issue reading the policy
+ language
+
+If the language was set but the policy wasn't, that could lead to
+a double free, as the value returned to the user was freed.
+---
+ lib/x509/x509_ext.c | 22 +++---
+ 1 file changed, 11 insertions(+), 11 deletions(-)
+
+diff --git a/lib/x509/x509_ext.c b/lib/x509/x509_ext.c
+index f974b0279..ed0ad1d14 100644
+--- a/lib/x509/x509_ext.c
 b/lib/x509/x509_ext.c
+@@ -1414,7 +1414,8 @@ int gnutls_x509_ext_import_proxy(const gnutls_datum_t * ext, int *pathlen,
+ {
+ 	ASN1_TYPE c2 = ASN1_TYPE_EMPTY;
+ 	int result;
+-	gnutls_datum_t value = { NULL, 0 };
++	gnutls_datum_t value1 = { NULL, 0 };
++	gnutls_datum_t value2 = { NULL, 0 };
+ 
+ 	if ((result = asn1_create_element
+ 	 (_gnutls_get_pkix(), "PKIX1.ProxyCertInfo",
+@@ -1444,20 +1445,18 @@ int gnutls_x509_ext_import_proxy(const gnutls_datum_t * ext, int *pathlen,
+ 	}
+ 
+ 	result = _gnutls_x509_read_value(c2, "proxyPolicy.policyLanguage",
+-	 );
++	 );
+ 	if (result < 0) {
+ 		gnutls_assert();
+ 		goto cleanup;
+ 	}
+ 
+ 	if (policyLanguage) {
+-		*policyLanguage = (char *)value.data;
+-	} else {
+-		gnutls_free(value.data);
+-		value.data = NULL;
++		*policyLanguage = (char *)value1.data;
++		value1.data = NULL;

Bug#849967: jessie-pu: package exim4/4.84.2-2+deb8u3

[Andreas Metzler]

Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu

Hello,

I (and Heiko from exim upstream) would like to fix #845569 in jessie.
sid/testing already include the fix, it was part of 4.88~RC6.

The issue is a memleak in the GnuTLS code, the patch is a towo line
change. Heiko has provided a very nice writeup in
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=845569#20

thanks, cu Andreas

-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru exim4-4.84.2/debian/changelog exim4-4.84.2/debian/changelog
--- exim4-4.84.2/debian/changelog   2016-12-22 12:17:01.0 +0100
+++ exim4-4.84.2/debian/changelog   2017-01-02 19:42:06.0 +0100
@@ -1,3 +1,11 @@
+exim4 (4.84.2-2+deb8u3) jessie; urgency=medium
+
+  * 94_Fix-memory-leak-on-Gnu-TLS-close.patch from upstream exim-4_84_2+fixes
+branch: Fix GnuTLS memory leak. (Thanks, Heiko Schlittermann!)
+Closes: #845569
+
+ -- Andreas Metzler <ametz...@debian.org>  Mon, 02 Jan 2017 19:18:05 +0100
+
 exim4 (4.84.2-2+deb8u2) jessie-security; urgency=high
 
   * Non-maintainer upload by the Security Team.
diff -Nru exim4-4.84.2/debian/patches/94_Fix-memory-leak-on-Gnu-TLS-close.patch 
exim4-4.84.2/debian/patches/94_Fix-memory-leak-on-Gnu-TLS-close.patch
--- exim4-4.84.2/debian/patches/94_Fix-memory-leak-on-Gnu-TLS-close.patch   
1970-01-01 01:00:00.0 +0100
+++ exim4-4.84.2/debian/patches/94_Fix-memory-leak-on-Gnu-TLS-close.patch   
2016-12-31 17:46:00.0 +0100
@@ -0,0 +1,52 @@
+From 867e8fe25dbfb1e31493488ad695bde55b890397 Mon Sep 17 00:00:00 2001
+From: "Heiko Schlittermann (HS12-RIPE)" <h...@schlittermann.de>
+Date: Wed, 23 Nov 2016 12:02:26 +0100
+Subject: [PATCH] Fix memory leak on (Gnu)TLS close.
+
+This leak doesn't show up under normal operation, as the process
+normally dies right after closing the session.
+
+But during callout repetitive TLS sessions are opened and closed from
+the same process (the process receiving the message). Depending on
+the amount of RAM and the number of callouts the same process does,
+this may be a problem. (On an amd64 machine with 4GB RAM, at about 1000
+recipients the memory is exhausted.)
+
+(cherry picked from commit ed62aae3051c9a713d35c8ae516fbd193d1401ba)
+---
+ src/tls-gnu.c | 5 +
+ 1 file changed, 5 insertions(+)
+
+diff --git a/src/tls-gnu.c b/src/tls-gnu.c
+index 61ed0e81..670f8cbc 100644
+--- a/src/tls-gnu.c
 b/src/tls-gnu.c
+@@ -1729,6 +1729,7 @@ if (rc != GNUTLS_E_SUCCESS)
+ 
+   if (!sigalrm_seen)
+ {
++gnutls_certificate_free_credentials(state->x509_cred);
+ (void)fclose(smtp_out);
+ (void)fclose(smtp_in);
+ }
+@@ -2014,6 +2015,8 @@ if (shutdown)
+   }
+ 
+ gnutls_deinit(state->session);
++gnutls_certificate_free_credentials(state->x509_cred);
++
+ 
+ state->tlsp->active = -1;
+ memcpy(state, _gnutls_state_init, sizeof(exim_gnutls_state_init));
+@@ -2074,6 +2077,8 @@ if (state->xfer_buffer_lwm >= state->xfer_buffer_hwm)
+ receive_smtp_buffered = smtp_buffered;
+ 
+ gnutls_deinit(state->session);
++gnutls_certificate_free_credentials(state->x509_cred);
++
+ state->session = NULL;
+ state->tlsp->active = -1;
+ state->tlsp->bits = 0;
+-- 
+2.11.0
+
diff -Nru exim4-4.84.2/debian/patches/series exim4-4.84.2/debian/patches/series
--- exim4-4.84.2/debian/patches/series  2016-12-22 12:17:01.0 +0100
+++ exim4-4.84.2/debian/patches/series  2016-12-31 17:45:15.0 +0100
@@ -23,3 +23,4 @@
 91_Expansions-Fix-crash-in-crypteq-On-OpenBSD-a-bad-sec.patch
 92_CVE-2016-1238.diff
 93_CVE-2016-9963-Fix-DKIM-information-leakage.patch
+94_Fix-memory-leak-on-Gnu-TLS-close.patch


signature.asc
Description: PGP signature

Bug#849436: unblock: exim4/4.88~RC6-2

[Andreas Metzler]

d5160d1c609170bdf2 d65072ddeb66cc7ad6950e23e0ea5d2ea76f9015+}
Version: [-4.88~RC6-1-] {+4.88~RC6-2+}

Control files of package exim4-dev: lines which differ (wdiff format)
-
Version: [-4.88~RC6-1-] {+4.88~RC6-2+}

Control files of package eximon4: lines which differ (wdiff format)
---
Version: [-4.88~RC6-1-] {+4.88~RC6-2+}
diff -Nru exim4-4.88~RC6/debian/changelog exim4-4.88~RC6/debian/changelog
--- exim4-4.88~RC6/debian/changelog	2016-12-08 07:19:18.0 +0100
+++ exim4-4.88~RC6/debian/changelog	2016-12-22 16:50:21.0 +0100
@@ -1,3 +1,15 @@
+exim4 (4.88~RC6-2) unstable; urgency=high
+
+  * Add macro IGNORE_SMTP_LINE_LENGTH_LIMIT to allow disabling the SMTP DATA
+physical line limit check for both for SMTP DATA ACL and remote_smtp*
+transports. Closes: #828801
+Also update corresponding NEWS entry.
+  * [lintian] debian/changelog: s/lenght/length/
+  * Pull 75_Fix-DKIM-information-leakage.patch from upstream GIT, fixing DKIM
+information leakage issue CVE-2016-9963.
+
+ -- Andreas Metzler <ametz...@debian.org>  Thu, 22 Dec 2016 16:50:21 +0100
+
 exim4 (4.88~RC6-1) unstable; urgency=low
 
   * New upstream version.
@@ -109,7 +121,7 @@
   expansion. https://bugs.exim.org/show_bug.cgi?id=165
   * Copy information message on rejecting overlong lines in data ACL from
 upstream example configuration. Closes: #823418
-  * Add NEWS entry on line-lenght-limit introduced in 4.87~RC1-1.
+  * Add NEWS entry on line-length-limit introduced in 4.87~RC1-1.
     Closes: 821830
 
  -- Andreas Metzler <ametz...@debian.org>  Sun, 08 May 2016 14:03:10 +0200
@@ -3805,7 +3817,7 @@
 - Supports CRL (Certificate Revocation List) (Closes: #229063)
 - exim_dbmbuild does not crash on _very_ long RHS values.
   (Closes: #231597)
-- route_list does not use a fixed lenght buffer anymore. (Closes: #231979)
+- route_list does not use a fixed length buffer anymore. (Closes: #231979)
 - An empty tls_verify_certificates file is correctly interpreted as empty
   list instead of breaking TLS. (Closes: #236478)
   * Korean translation of debconf templates by Changwoo Ryu (Closes: #241499)
diff -Nru exim4-4.88~RC6/debian/debconf/conf.d/acl/40_exim4-config_check_data exim4-4.88~RC6/debian/debconf/conf.d/acl/40_exim4-config_check_data
--- exim4-4.88~RC6/debian/debconf/conf.d/acl/40_exim4-config_check_data	2016-09-25 14:46:29.0 +0200
+++ exim4-4.88~RC6/debian/debconf/conf.d/acl/40_exim4-config_check_data	2016-12-18 13:59:15.0 +0100
@@ -11,9 +11,11 @@
   # Deny if the message contains an overlong line.  Per the standards
   # we should never receive one such via SMTP.
   #
+  .ifndef IGNORE_SMTP_LINE_LENGTH_LIMIT
   denymessage= maximum allowed line length is 998 octets, \
got $max_received_linelength
   condition  = ${if > {$max_received_linelength}{998}}
+  .endif
 
   # Deny unless the address list headers are syntactically correct.
   #
diff -Nru exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp
--- exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp	2016-09-25 14:46:29.0 +0200
+++ exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp	2016-12-18 13:59:52.0 +0100
@@ -9,7 +9,9 @@
 remote_smtp:
   debug_print = "T: remote_smtp for $local_part@$domain"
   driver = smtp
+.ifndef IGNORE_SMTP_LINE_LENGTH_LIMIT
   message_size_limit = ${if > {$max_received_linelength}{998} {1}{0}}
+.endif
 .ifdef REMOTE_SMTP_HOSTS_AVOID_TLS
   hosts_avoid_tls = REMOTE_SMTP_HOSTS_AVOID_TLS
 .endif
diff -Nru exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp_smarthost exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp_smarthost
--- exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp_smarthost	2016-09-25 14:46:29.0 +0200
+++ exim4-4.88~RC6/debian/debconf/conf.d/transport/30_exim4-config_remote_smtp_smarthost	2016-12-18 14:00:13.0 +0100
@@ -12,7 +12,9 @@
 remote_smtp_smarthost:
   debug_print = "T: remote_smtp_smarthost for $local_part@$domain"
   driver = smtp
+.ifndef IGNORE_SMTP_LINE_LENGTH_LIMIT
   message_size_limit = ${if > {$max_received_linelength}{998} {1}{0}}
+.endif
   hosts_try_auth = <; ${if exists{CONFDIR/passwd.client} \
 {\
 ${lookup{$host}nwildlsearch{CONFDIR/passwd.client}{$host_address}}\
diff -Nru exim4-4.88~RC6/debian/NEWS exim4-4.88~RC6/debian/NEWS
--- exim4-4.88~RC6/debian/NEWS	2016-09-25 14:46:29.0 +0200
+++ exim4-4.88~RC6/debian/NEWS	2016-12-18 14:04:32.0 +0100
@@ -1,9 +1,11 @@
 exim4 (4.87-3) unstable; urgency=medium
 
-  Starting with 4.87~RC1-1 exim will not accept messages with physical lines
-  long

Bug#849329: unblock: gnutls28/3.5.7-3

[Andreas Metzler]

 libgnutls-dane0: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.7-2),-] {+3.5.7-3),+} libc6 (>= 2.14), libunbound2 (>= 1.4.1)
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls-dane0-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-4942f0c0688463070e6410365999f7a60d5bde23-] {+c1ead7f61001838e6d88ff1cd74ac74e22c469f4+}
Depends: libgnutls-dane0 (= [-3.5.7-2)-] {+3.5.7-3)+}
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls-openssl27: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.7-2),-] {+3.5.7-3),+} libc6 (>= 2.14)
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls-openssl27-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-9d6f39cb57ee78768fb728e590d19669272f0816-] {+c5ed28d817ac7aaf9d6a0aa028f34f13e57f7a45+}
Depends: libgnutls-openssl27 (= [-3.5.7-2)-] {+3.5.7-3)+}
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls28-dev: lines which differ (wdiff format)
---
Depends: libgnutls30 (= [-3.5.7-2),-] {+3.5.7-3),+} libgnutls-openssl27 (= [-3.5.7-2),-] {+3.5.7-3),+} libgnutlsxx28 (= [-3.5.7-2),-] {+3.5.7-3),+} libgnutls-dane0 (= [-3.5.7-2),-] {+3.5.7-3),+} nettle-dev, libc6-dev | libc-dev, zlib1g-dev, libtasn1-6-dev, libp11-kit-dev, libidn11-dev (>= 1.31)
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls30: lines which differ (wdiff format)
---
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutls30-dbgsym: lines which differ (wdiff format)
--
Build-Ids: [-2549b7cc772d8fd074de0be00f0619db53bee1f1-] {+9addeb34b9f349ee50037cd28d46fc5c9112c6fe+}
Depends: libgnutls30 (= [-3.5.7-2)-] {+3.5.7-3)+}
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutlsxx28: lines which differ (wdiff format)
-
Depends: libgnutls30 (= [-3.5.7-2),-] {+3.5.7-3),+} libc6 (>= 2.4), libgcc1 (>= 1:3.0), libstdc++6 (>= 5)
Version: [-3.5.7-2-] {+3.5.7-3+}

Control files of package libgnutlsxx28-dbgsym: lines which differ (wdiff format)

Build-Ids: [-cdb980046cd934ff2b0fedb5235e56484dcfadcd-] {+0692627b5d607063eb71903a721233f5901066e9+}
Depends: libgnutlsxx28 (= [-3.5.7-2)-] {+3.5.7-3)+}
Version: [-3.5.7-2-] {+3.5.7-3+}
diff -Nru gnutls28-3.5.7/debian/changelog gnutls28-3.5.7/debian/changelog
--- gnutls28-3.5.7/debian/changelog	2016-12-09 18:10:53.0 +0100
+++ gnutls28-3.5.7/debian/changelog	2016-12-20 18:47:13.0 +0100
@@ -1,3 +1,14 @@
+gnutls28 (3.5.7-3) unstable; urgency=medium
+
+  * 35_01_pkcs8-ensure-that-the-correct-error-code-is-returned.patch,
+35_02_tests-added-test-for-PKCS-8-encrypted-key-decoding.patch from
+upstream 3.5 branch: Ensure that GNUTLS_E_DECRYPTION_FAIL will be returned
+by PKCS#8 decryption functions when an invalid key is provided. This
+addresses regression on decrypting certain PKCS#8 keys.
+Closes: #848905
+
+ -- Andreas Metzler <ametz...@debian.org>  Tue, 20 Dec 2016 18:47:13 +0100
+
 gnutls28 (3.5.7-2) unstable; urgency=medium
 
   * Upload to unstable.
diff -Nru gnutls28-3.5.7/debian/patches/35_01_pkcs8-ensure-that-the-correct-error-code-is-returned.patch gnutls28-3.5.7/debian/patches/35_01_pkcs8-ensure-that-the-correct-error-code-is-returned.patch
--- gnutls28-3.5.7/debian/patches/35_01_pkcs8-ensure-that-the-correct-error-code-is-returned.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.5.7/debian/patches/35_01_pkcs8-ensure-that-the-correct-error-code-is-returned.patch	2016-12-20 18:39:09.0 +0100
@@ -0,0 +1,25 @@
+From e62aaf4bfaf1a4280db23d9729c2d7fa0fdf97e5 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@redhat.com>
+Date: Tue, 13 Dec 2016 11:27:38 +0100
+Subject: [PATCH 1/3] pkcs8: ensure that the correct error code is returned on
+ decryption failure
+
+---
+ lib/x509/privkey_pkcs8.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/lib/x509/privkey_pkcs8.c b/lib/x509/privkey_pkcs8.c
+index 74bb466c6..0094a83a5 100644
+--- a/lib/x509/privkey_pkcs8.c
 b/lib/x509/privkey_pkcs8.c
+@@ -711,6 +711,7 @@ static int pkcs8_key_decrypt(const gnutls_datum_t * raw_key,
+ 			 _params, _params, );
+ 	if (result < 0) {
+ 		gnutls_assert();
++		result = GNUTLS_E_DECRYPTION_FAILED;
+ 		goto error;
+ 	}
+ 
+-- 
+2.11.0
+
diff -Nru gnutls28-3.5.7/debian/patches/35_02_tests

Bug#840191: jessie-pu: package gnutls28/3.3.8-6+deb8u4

[Andreas Metzler]

On 2016-10-31 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> On Sun, 2016-10-30 at 07:46 +0100, Andreas Metzler wrote:
[...]
>> I think it makes sense to  add the GnuTLS patch for compatibitlity with
>> CVE-2016-6489-patched nettle. (832983).

> jessie's nettle doesn't appear to have been updated for that issue, but
> I guess it still makes sense to include this for partial upgrades.

Thank you, uploaded.

The update for CVE-2016-6489 is currently being worked on and
a fixed gnutls version needs to be available before.

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'

Bug#840191: jessie-pu: package gnutls28/3.3.8-6+deb8u4

[Andreas Metzler]

On 2016-10-09 Salvatore Bonaccorso <car...@debian.org> wrote:
[...]
> Hi Stable Release Managers,

> X-Debbugs-CC'ed Andreas Metzler.

> gnutls28 in jessie is affected by CVE-2016-7444, GNUTLS-SA-2016-3,
> having a flaw in the OCSP certificate check. This was fixed upstream
> and included in unstable with 3.5.3-4 but would not warrant a DSA.

> Attached is proposed debdiff for jessie. Would it be acceptable for an
> upcoming point release?
[...]

I think it makes sense to  add the GnuTLS patch for compatibitlity with
CVE-2016-6489-patched nettle. (832983).

cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'
diff -Nru gnutls28-3.3.8/debian/changelog gnutls28-3.3.8/debian/changelog
--- gnutls28-3.3.8/debian/changelog	2015-08-14 18:29:51.0 +0200
+++ gnutls28-3.3.8/debian/changelog	2016-10-30 07:39:11.0 +0100
@@ -1,3 +1,16 @@
+gnutls28 (3.3.8-6+deb8u4) jessie; urgency=medium
+
+  [ Salvatore Bonaccorso ]
+  * CVE-2016-7444: Incorrect certificate validation when using OCSP responses
+(GNUTLS-SA-2016-3). See #840191.
+
+  [ Andreas Metzler ]
+  * Cherry pick 53_nettle-use-rsa_-_key_prepare-on-key-import.patch
+from upstream GIT, which should allow gnutls continue to work with
+CVE-2016-6489-patched nettle. See #832983.
+
+ -- Andreas Metzler <ametz...@debian.org>  Sun, 30 Oct 2016 07:39:08 +0100
+
 gnutls28 (3.3.8-6+deb8u3) jessie; urgency=medium
 
   * Pull 50_Handle-zero-length-plaintext-for-VIA-PadLock-functio.patch from
diff -Nru gnutls28-3.3.8/debian/patches/52_CVE-2016-7444_ocsp-corrected-the-comparison-of-the-serial-size-in-.patch gnutls28-3.3.8/debian/patches/52_CVE-2016-7444_ocsp-corrected-the-comparison-of-the-serial-size-in-.patch
--- gnutls28-3.3.8/debian/patches/52_CVE-2016-7444_ocsp-corrected-the-comparison-of-the-serial-size-in-.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.3.8/debian/patches/52_CVE-2016-7444_ocsp-corrected-the-comparison-of-the-serial-size-in-.patch	2016-10-30 07:01:40.0 +0100
@@ -0,0 +1,24 @@
+From 964632f37dfdfb914ebc5e49db4fa29af35b1de9 Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@gnutls.org>
+Date: Sat, 27 Aug 2016 17:00:22 +0200
+Subject: [PATCH] ocsp: corrected the comparison of the serial size in OCSP
+ response
+
+Previously the OCSP certificate check wouldn't verify the serial length
+and could succeed in cases it shouldn't.
+
+Reported by Stefan Buehler.
+---
+ lib/x509/ocsp.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+--- a/lib/x509/ocsp.c
 b/lib/x509/ocsp.c
+@@ -1251,6 +1251,7 @@ gnutls_ocsp_resp_check_crt(gnutls_ocsp_r
+ 		gnutls_assert();
+ 		goto cleanup;
+ 	}
++	cserial.size = t;
+ 
+ 	if (rserial.size != cserial.size
+ 	|| memcmp(cserial.data, rserial.data, rserial.size) != 0) {
diff -Nru gnutls28-3.3.8/debian/patches/53_nettle-use-rsa_-_key_prepare-on-key-import.patch gnutls28-3.3.8/debian/patches/53_nettle-use-rsa_-_key_prepare-on-key-import.patch
--- gnutls28-3.3.8/debian/patches/53_nettle-use-rsa_-_key_prepare-on-key-import.patch	1970-01-01 01:00:00.0 +0100
+++ gnutls28-3.3.8/debian/patches/53_nettle-use-rsa_-_key_prepare-on-key-import.patch	2016-10-30 07:10:31.0 +0100
@@ -0,0 +1,152 @@
+From 186dc9c2012003587a38d7f4d03edd8da5fe989f Mon Sep 17 00:00:00 2001
+From: Nikos Mavrogiannopoulos <n...@gnutls.org>
+Date: Sun, 7 Aug 2016 12:06:39 +0200
+Subject: [PATCH] nettle: use rsa_*_key_prepare on key import
+
+Previously we calculated the size of the key directly, but
+by using the rsa_*_key_prepare we benefit from any checks that
+may be introduced in the future. Specifically any checks for invalid
+public keys (e.g., keys that may crash the underlying gmp functions).
+
+This patch avoids calling rsa_private_key_prepare every time we construct
+a nettle private key struct, because this function requires a bigint
+multiplication. We call that function once on private key import.
+---
+ lib/nettle/pk.c | 62 +++--
+ 1 file changed, 52 insertions(+), 10 deletions(-)
+
+diff --git a/lib/nettle/pk.c b/lib/nettle/pk.c
+index 2fab308..b41ebfb 100644
+--- a/lib/nettle/pk.c
 b/lib/nettle/pk.c
+@@ -98,18 +98,24 @@ _rsa_params_to_privkey(const gnutls_pk_params_st * pk_params,
+ 	memcpy(priv->c, pk_params->params[5], SIZEOF_MPZT);
+ 	memcpy(priv->a, pk_params->params[6], SIZEOF_MPZT);
+ 	memcpy(priv->b, pk_params->params[7], SIZEOF_MPZT);
++	/* we do not rsa_private_key_prepare() because it involves a multiplication.
++	 * we call it once when we import the parameters */
+ 	priv->size =
+ 	nettle_mpz_sizeinbase_256_u(TOMPZ
+ 	(pk_params->params[RSA_MODULUS]));
+ }
+ 
+-static void
++/* returns a negative value on invalid pubkey */
++static int
+ _rsa_params_to_pubkey(const gnutls_pk_params_st * pk_params,
+ 		  struct rsa_public_key *pub)
+ {
+ 	memcpy(p

Bug#827111: jessie-pu: package exim4/4.84.2-2

[Andreas Metzler]

On 2016-07-26 Salvatore Bonaccorso  wrote:
> On Mon, Jul 25, 2016 at 08:50:47PM +0200, Salvatore Bonaccorso wrote:
>> On Mon, Jul 25, 2016 at 07:28:33PM +0100, Adam D. Barratt wrote:
>>> On Mon, 2016-07-25 at 20:14 +0200, Salvatore Bonaccorso wrote:
[...]
 Since we claimed 4.84.2-2+deb8u1 in the DSA, would it help if we just
 redo the update, push the packages? (without further announce, since
 that was the claimed version)?
 
 Attached how that would look like with debdiff against 4.84.2-2.
>>> 
>>> If you're happy to do so, that looks good to me; thanks.
>> 
>> I just have uploaded 4.84.2-2+deb8u1 and will dak install the builds.
>> Hope no user will complain.

> FTR, this has been done. 

Thank you!

Bug#827111: jessie-pu: package exim4/4.84.2-2

[Andreas Metzler]

On 2016-06-17 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> Control: tags -1 + pending

> On Thu, 2016-06-16 at 18:38 +0200, Andreas Metzler wrote:
> > On 2016-06-12 "Adam D. Barratt" <a...@adam-barratt.org.uk> wrote:
> > > Control: tags -1 + confirmed
> > 
> > > On Sun, 2016-06-12 at 14:06 +0200, Andreas Metzler wrote:
> > >> I would like to update exim4 in jessie with the following changes:
[...]
> Flagged for acceptance.


Hello,

now we have 4.84.2-1+deb8u1 in stable security and 4.84.2-2 in spu would
overwrite it at the next stable release. How do I fix this properly?

a) Redo 4.84.2-2 with 4.84.2-1+deb8u1 merged in
b) Release 4.84.2-3 with 4.84.2-1+deb8u1 merged in

TIA, cu Andreas
-- 
`What a good friend you are to him, Dr. Maturin. His other friends are
so grateful to you.'
`I sew his ears on from time to time, sure'