]> git.ipfire.org Git - thirdparty/gnutls.git/log
thirdparty/gnutls.git
7 years agoMerge branch 'tmp-reset-after-free' into 'master'
Tim Rühsen [Sat, 23 Feb 2019 22:21:09 +0000 (22:21 +0000)] 
Merge branch 'tmp-reset-after-free' into 'master'

Automatically NULLify after gnutls_free()

See merge request gnutls/gnutls!923

7 years agoMerge branch 'tmp-cleanup-cert' into 'master'
Tim Rühsen [Sat, 23 Feb 2019 22:20:21 +0000 (22:20 +0000)] 
Merge branch 'tmp-cleanup-cert' into 'master'

Cleanup lib/auth/cert.c as suggested by cppcheck

See merge request gnutls/gnutls!924

7 years ago_gnutls_recv_handshake: added explicit sanity checks
Nikos Mavrogiannopoulos [Sat, 23 Feb 2019 20:19:06 +0000 (21:19 +0100)] 
_gnutls_recv_handshake: added explicit sanity checks

Although, this function acts on the message provided as expected and thus
it should never call a message parsing function on unexpected
messages, we make a more explicit sanity check. This unifies the
sanity checks existing within the involved functions.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agognutls_x509_crt_init: Fix dereference of NULL pointer
Tim Rühsen [Tue, 12 Feb 2019 14:20:23 +0000 (15:20 +0100)] 
gnutls_x509_crt_init: Fix dereference of NULL pointer

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoRemove redundant resets of variables after free()
Tim Rühsen [Tue, 12 Feb 2019 14:14:07 +0000 (15:14 +0100)] 
Remove redundant resets of variables after free()

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoAutomatically NULLify after gnutls_free()
Tim Rühsen [Tue, 12 Feb 2019 14:09:11 +0000 (15:09 +0100)] 
Automatically NULLify after gnutls_free()

This method prevents direct use-after-free and
double-free issues.

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-downgrade-sentinel' into 'master'
Daiki Ueno [Fri, 22 Feb 2019 11:42:27 +0000 (11:42 +0000)] 
Merge branch 'tmp-downgrade-sentinel' into 'master'

handshake: defer setting downgrade sentinel until version is selected

Closes #689

See merge request gnutls/gnutls!918

7 years agoMerge branch 'use_topendir_fixed' into 'master'
Tim Rühsen [Fri, 22 Feb 2019 10:23:31 +0000 (10:23 +0000)] 
Merge branch 'use_topendir_fixed' into 'master'

Re-introduce topendir on Windows with Unicode support

See merge request gnutls/gnutls!932

7 years agotlsfuzzer: update to the latest upstream for downgrade protection tests
Daiki Ueno [Tue, 19 Feb 2019 12:56:35 +0000 (13:56 +0100)] 
tlsfuzzer: update to the latest upstream for downgrade protection tests

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoext/supported_versions: regenerate server random
Daiki Ueno [Sat, 9 Feb 2019 09:26:56 +0000 (10:26 +0100)] 
ext/supported_versions: regenerate server random

This adds a call to _gnutls_gen_server_random() in handling the
"supported_versions" extension, so that the TLS 1.3 downgrade sentinel
is set only when the earlier versions are selected.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoUpdate ax_code_coverage.m4 to latest release of autoconf-archive
Tim Rühsen [Tue, 5 Feb 2019 10:01:20 +0000 (11:01 +0100)] 
Update ax_code_coverage.m4 to latest release of autoconf-archive

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'thread_local_msvc_fixed' into 'master'
Nikos Mavrogiannopoulos [Thu, 21 Feb 2019 18:13:48 +0000 (18:13 +0000)] 
Merge branch 'thread_local_msvc_fixed' into 'master'

lib: Provide _Thread_local on MSVC

See merge request gnutls/gnutls!933

7 years agolib: x509: Minor directory browsing simplification
Hugo Beauzée-Luyssen [Thu, 21 Feb 2019 13:49:36 +0000 (14:49 +0100)] 
lib: x509: Minor directory browsing simplification

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
7 years agoRevert "Revert "verify-high2: Fix cert dir iteration on Win32""
Hugo Beauzée-Luyssen [Mon, 18 Feb 2019 13:41:56 +0000 (14:41 +0100)] 
Revert "Revert "verify-high2: Fix cert dir iteration on Win32""

This reverts commit 681330882da19099eea360fab141cab937c45677.

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
This revert also contains the fix to the original commit (invalid
utf8->utf16 conversion) and a minor simplification of the _treaddir loop.

7 years agoiconv: Allow _gnutls_utf8_to_ucs2 to output little endian
Hugo Beauzée-Luyssen [Mon, 18 Feb 2019 16:12:54 +0000 (17:12 +0100)] 
iconv: Allow _gnutls_utf8_to_ucs2 to output little endian

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
7 years agolib: Provide _Thread_local on MSVC
Hugo Beauzée-Luyssen [Mon, 18 Feb 2019 08:37:04 +0000 (09:37 +0100)] 
lib: Provide _Thread_local on MSVC

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
7 years agoMerge branch 'tmp-get-source-branch' into 'master'
Nikos Mavrogiannopoulos [Tue, 19 Feb 2019 07:20:17 +0000 (07:20 +0000)] 
Merge branch 'tmp-get-source-branch' into 'master'

check_if_signed: Get source branch if not set

See merge request gnutls/gnutls!930

7 years agoAdd test for starttls XMPP
Tim Rühsen [Mon, 18 Feb 2019 20:38:38 +0000 (21:38 +0100)] 
Add test for starttls XMPP

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agognutls-cli: Fix --starttls-proto=xmpp
Tim Rühsen [Wed, 6 Feb 2019 10:30:06 +0000 (11:30 +0100)] 
gnutls-cli: Fix --starttls-proto=xmpp

Fixes two issues with gnutls-cli --starttls-proto=xmpp:
1. Print 'Timeout' on timeout instead of random errno message
2. Do not wait for linefeed when using XMPP (XML)

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agocheck_if_signed: Get source branch if not set
Tim Rühsen [Mon, 18 Feb 2019 14:38:56 +0000 (15:38 +0100)] 
check_if_signed: Get source branch if not set

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-fix-record-size-limit-resumption' into 'master'
Daiki Ueno [Mon, 18 Feb 2019 12:10:11 +0000 (12:10 +0000)] 
Merge branch 'tmp-fix-record-size-limit-resumption' into 'master'

Fix issues in record_size_limit extension handling

See merge request gnutls/gnutls!879

7 years agoMerge branch 'gnutls-703' into 'master'
Tim Rühsen [Thu, 14 Feb 2019 19:31:14 +0000 (19:31 +0000)] 
Merge branch 'gnutls-703' into 'master'

tests: wrap ADD_SYSCALL for getrandom in test for SYS_getrandom

Closes #703

See merge request gnutls/gnutls!926

7 years agotests: wrap ADD_SYSCALL for getrandom in test for SYS_getrandom
R. Andrew Bailey [Thu, 14 Feb 2019 14:38:33 +0000 (09:38 -0500)] 
tests: wrap ADD_SYSCALL for getrandom in test for SYS_getrandom

Signed-off-by: R. Andrew Bailey <bailey@akamai.com>
7 years agognutls_record_set_max_size: make it work on server side
Daiki Ueno [Fri, 8 Feb 2019 13:46:33 +0000 (14:46 +0100)] 
gnutls_record_set_max_size: make it work on server side

The record_size_limit extension can also be specified by the server to
indicate the maximum plaintext.  Also add test cases for asymmetric
settings between server and client.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agotlsfuzzer: update to the latest upstream for record_size_limit test
Daiki Ueno [Thu, 31 Jan 2019 12:39:35 +0000 (13:39 +0100)] 
tlsfuzzer: update to the latest upstream for record_size_limit test

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoMerge branch 'tmp-macosx-fix' into 'master'
Tim Rühsen [Thu, 14 Feb 2019 14:55:56 +0000 (14:55 +0000)] 
Merge branch 'tmp-macosx-fix' into 'master'

bootstrap.conf: do not override GNULIB_SRCDIR

See merge request gnutls/gnutls!925

7 years agoext/record_size_limit: account for content type octet in TLS 1.3
Daiki Ueno [Fri, 8 Feb 2019 12:22:13 +0000 (13:22 +0100)] 
ext/record_size_limit: account for content type octet in TLS 1.3

In TLS 1.3, the protocol maximum of plaintext size is 2^14+1, while
it is 2^14 in TLS 1.2.  To accommodate that, this introduces the
following invariant:
- when the maximum is set by the user with
  gnutls_record_set_max_size(), store it as is.  The value range is
  [511, 16834].
- when the maximum is negotiated through record_size_limit extension,
  it can be [512, 16385].  In TLS 1.3, subtract by 1 to fit in [511,
  16384].

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agodecrypt_packet_tls13: add check for max plaintext size
Daiki Ueno [Thu, 31 Jan 2019 15:56:55 +0000 (16:56 +0100)] 
decrypt_packet_tls13: add check for max plaintext size

There is check in _gnutls_recv_in_buffers already, but for TLS 1.3 we
need to take account of the padding.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agorecord: reject too large plaintext after decryption
Daiki Ueno [Fri, 25 Jan 2019 16:00:44 +0000 (17:00 +0100)] 
record: reject too large plaintext after decryption

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoconstate: reset max_record_recv_size upon renegotiation
Daiki Ueno [Wed, 30 Jan 2019 15:45:08 +0000 (16:45 +0100)] 
constate: reset max_record_recv_size upon renegotiation

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agosession_pack: reset max_record_recv_size when packing
Daiki Ueno [Wed, 30 Jan 2019 09:21:07 +0000 (10:21 +0100)] 
session_pack: reset max_record_recv_size when packing

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoext/record_size_limit: don't confuse with negotiated/user-supplied maximum
Daiki Ueno [Thu, 17 Jan 2019 10:53:35 +0000 (11:53 +0100)] 
ext/record_size_limit: don't confuse with negotiated/user-supplied maximum

As documented in gnutls_int.h, max_record_send_size is for tracking
the user-supplied maximum, while max_record_recv_size for the
protocol negotiated maximum.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoext/max_record: server shouldn't send it with record_size_limit
Daiki Ueno [Sun, 20 Jan 2019 08:18:21 +0000 (09:18 +0100)] 
ext/max_record: server shouldn't send it with record_size_limit

Otherwise, the connection will be disconnected by the client, as
suggested in RFC: A client MUST treat receipt of both
"max_fragment_length" and "record_size_limit" as a fatal error, and it
SHOULD generate an "illegal_parameter" alert.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years ago_gnutls_hello_ext_is_present: don't ignore max_fragment_length
Daiki Ueno [Thu, 7 Feb 2019 15:28:52 +0000 (16:28 +0100)] 
_gnutls_hello_ext_is_present: don't ignore max_fragment_length

The extension is assigned the internal ID 0.

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years ago.dir-locals.el: disable indent-tabs-mode in js-mode
Daiki Ueno [Fri, 25 Jan 2019 16:04:40 +0000 (17:04 +0100)] 
.dir-locals.el: disable indent-tabs-mode in js-mode

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agobootstrap.conf: do not override GNULIB_SRCDIR
Nikos Mavrogiannopoulos [Thu, 14 Feb 2019 12:01:34 +0000 (13:01 +0100)] 
bootstrap.conf: do not override GNULIB_SRCDIR

This was not set in all of our CI platforms, and was causing
issues in MacOSX.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoMerge branch 'tmp-fix-cert-params-checks' into 'master'
Nikos Mavrogiannopoulos [Thu, 14 Feb 2019 11:57:23 +0000 (11:57 +0000)] 
Merge branch 'tmp-fix-cert-params-checks' into 'master'

x509: corrected issue in the algorithm parameters comparison

Closes #698

See merge request gnutls/gnutls!921

7 years agoMerge branch 'tmp-fix-uninitialized2' into 'master'
Tim Rühsen [Thu, 14 Feb 2019 11:07:15 +0000 (11:07 +0000)] 
Merge branch 'tmp-fix-uninitialized2' into 'master'

Fix uninitialized warning in pkcs11.c

See merge request gnutls/gnutls!906

7 years agox509: corrected issue in the algorithm parameters comparison
Nikos Mavrogiannopoulos [Mon, 11 Feb 2019 08:18:46 +0000 (09:18 +0100)] 
x509: corrected issue in the algorithm parameters comparison

Each certificate has two fields to set the signature algorithm
and parameters used for the digital signature. One of the fields is
authenticated and the other is not. It is required from RFC5280 to
enforce the equality of these fields, but currently due to an issue
we wouldn't enforce the equality of the parameters fields. This
fix corrects the issue.

We also move an RSA-PSS certificate in chainverify that was relying
on invalid parameters, to this set of invalid certificates.

Resolves: #698

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agotests: added further checks for gnutls_pkcs11_token_get_info
Nikos Mavrogiannopoulos [Wed, 13 Feb 2019 07:54:59 +0000 (07:54 +0000)] 
tests: added further checks for gnutls_pkcs11_token_get_info

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoFix uninitialized warning in pkcs11.c
Tim Rühsen [Tue, 29 Jan 2019 15:10:59 +0000 (16:10 +0100)] 
Fix uninitialized warning in pkcs11.c

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoCleanup lib/auth/cert.c as suggested by cppcheck
Tim Rühsen [Wed, 13 Feb 2019 16:22:21 +0000 (17:22 +0100)] 
Cleanup lib/auth/cert.c as suggested by cppcheck

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-serv-args' into 'master'
Tim Rühsen [Tue, 12 Feb 2019 16:53:01 +0000 (16:53 +0000)] 
Merge branch 'tmp-serv-args' into 'master'

Fix 32bit overflow issue in src/serv-args.def

Closes #700

See merge request gnutls/gnutls!922

7 years agoFix 32bit overflow issue in src/serv-args.def
Tim Rühsen [Mon, 11 Feb 2019 09:41:47 +0000 (10:41 +0100)] 
Fix 32bit overflow issue in src/serv-args.def

Fixing this warning seen on 32bit architectures:

serv-args.c: In function 'doOptMaxearlydata':
serv-args.c:1431:14: warning: overflow in conversion from 'long long int' to 'long int' changes value from '4294967296' to '0' [-Woverflow]
         { 1, 4294967296 } };
              ^~~~~~~~~~

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-ssize-t' into 'master'
Tim Rühsen [Tue, 12 Feb 2019 11:34:22 +0000 (11:34 +0000)] 
Merge branch 'tmp-ssize-t' into 'master'

Remove typedef'ing ssize_t in gnutls.h

Closes #688

See merge request gnutls/gnutls!916

7 years agoRemove typedef'ing ssize_t in gnutls.h
Tim Rühsen [Fri, 8 Feb 2019 12:03:30 +0000 (13:03 +0100)] 
Remove typedef'ing ssize_t in gnutls.h

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-gnulib-pton' into 'master'
Tim Rühsen [Sun, 10 Feb 2019 15:46:55 +0000 (15:46 +0000)] 
Merge branch 'tmp-gnulib-pton' into 'master'

Use inet_pton() from gnulib

See merge request gnutls/gnutls!913

7 years agoUse inet_pton() from gnulib
Tim Rühsen [Wed, 6 Feb 2019 19:54:45 +0000 (20:54 +0100)] 
Use inet_pton() from gnulib

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-gitlab-triage' into 'master'
Tim Rühsen [Sun, 10 Feb 2019 07:26:01 +0000 (07:26 +0000)] 
Merge branch 'tmp-gitlab-triage' into 'master'

.triage-policies.yml: added [ci skip]

See merge request gnutls/gnutls!908

7 years agoMerge branch 'tmp-enforce-gnulib-rules-for-lib' into 'master'
Tim Rühsen [Sat, 9 Feb 2019 18:06:41 +0000 (18:06 +0000)] 
Merge branch 'tmp-enforce-gnulib-rules-for-lib' into 'master'

bootstrap: refuse to bootstrap if any new dependencies bring gnulib's network stack

See merge request gnutls/gnutls!919

7 years agobootstrap: refuse to bootstrap if any dependencies bring gnulib's network stack
Nikos Mavrogiannopoulos [Sat, 9 Feb 2019 09:52:29 +0000 (10:52 +0100)] 
bootstrap: refuse to bootstrap if any dependencies bring gnulib's network stack

If gnulib's network stack is brought (due to a dependency) in the library
it will make the library unusable to non-gnulib using applications. This
prevents windows applications for example to use gnutls, and so on. Even
more it is quite hard to catch that issue because our testsuite uses
gnulib as well. Instead we try to catch the these modules at import time.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoMerge branch 'tmp-key-usage' into 'master'
Nikos Mavrogiannopoulos [Fri, 8 Feb 2019 17:42:10 +0000 (17:42 +0000)] 
Merge branch 'tmp-key-usage' into 'master'

When negotiating TLS1.3 enforce certificate key usage

Closes #690

See merge request gnutls/gnutls!902

7 years agoMerge branch 'tmp-gnulib-ntop' into 'master'
Tim Rühsen [Fri, 8 Feb 2019 13:04:47 +0000 (13:04 +0000)] 
Merge branch 'tmp-gnulib-ntop' into 'master'

Use inet_ntop() from gnulib

See merge request gnutls/gnutls!912

7 years agoUse inet_ntop() from gnulib
Tim Rühsen [Wed, 6 Feb 2019 19:35:11 +0000 (20:35 +0100)] 
Use inet_ntop() from gnulib

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years ago_gnutls_gen_rawpk_crt: corrected the use of assert
Nikos Mavrogiannopoulos [Sun, 3 Feb 2019 11:18:30 +0000 (12:18 +0100)] 
_gnutls_gen_rawpk_crt: corrected the use of assert

The API could return 0 or 1 matching certificates. The case of zero
can only happen in client side.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoraw public keys: apply the key usage bits the same way as X.509
Nikos Mavrogiannopoulos [Sun, 3 Feb 2019 07:47:50 +0000 (08:47 +0100)] 
raw public keys: apply the key usage bits the same way as X.509

That is, we require a signing certificate when negotiating
TLS1.3, or when sending a client certificate (on all cases).

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoFallback to TLS 1.2 when incompatible with signature certs are provided
Nikos Mavrogiannopoulos [Sat, 2 Feb 2019 08:13:40 +0000 (09:13 +0100)] 
Fallback to TLS 1.2 when incompatible with signature certs are provided

This only takes into account certificates in the credentials structure.
If certificates are provided in a callback, these must be checked by
the provider. For that we assume that the credentials structure is
filled when associated with a session; if not then the fallback mechanism
will not work and the handshake will fail.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoEnforce the certificate key usage restrictions on all cases
Nikos Mavrogiannopoulos [Sat, 2 Feb 2019 06:10:10 +0000 (07:10 +0100)] 
Enforce the certificate key usage restrictions on all cases

That is, we require a signing certificate when negotiating
TLS1.3, or when sending a client certificate (on all cases).

Before we would not perform any checks under TLS1.3 or when client
certificates are sent, assuming that the certificates used will always
be signing ones. However if the user sets up incorrectly a decryption
certificate we would use it for signing. This fix makes sure that an
error is returned early when these scenarios are detected.

Resolves: #690

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoFetch OSS-Fuzz corpora much faster [skip ci]
Tim Rühsen [Mon, 21 Jan 2019 11:54:58 +0000 (12:54 +0100)] 
Fetch OSS-Fuzz corpora much faster [skip ci]

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years ago.triage-policies.yml: added [ci skip]
Nikos Mavrogiannopoulos [Mon, 4 Feb 2019 14:10:19 +0000 (15:10 +0100)] 
.triage-policies.yml: added [ci skip]

This adds a set of policies regarding issues and merge requests
to be enforced by the gitlab-triage bot. That is:
 - Issues without any label for more than a month are marked
   with needs attention label
 - Issues with needinfo label are closed if they are not updated
   within a month
 - Merge requests marked as WIP with no update within 5 months
   are closed.

These rules are not enforced automatically; we have to schedule
a run of the gitlab-triage bot.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
7 years agoMerge branch 'fix-nettle-check' into 'master'
Nikos Mavrogiannopoulos [Mon, 4 Feb 2019 13:16:11 +0000 (13:16 +0000)] 
Merge branch 'fix-nettle-check' into 'master'

build: pass NETTLE_LIBS together with HOGWEED_LIBS

See merge request gnutls/gnutls!903

7 years agoMerge branch 'fix-pkcs11-mechs' into 'master'
Nikos Mavrogiannopoulos [Mon, 4 Feb 2019 13:11:56 +0000 (13:11 +0000)] 
Merge branch 'fix-pkcs11-mechs' into 'master'

build: do not generate mech-list.h if p11-kit is not available

See merge request gnutls/gnutls!904

7 years agobuild: do not generate mech-list.h if p11-kit is not available
Dmitry Eremin-Solenikov [Sat, 2 Feb 2019 14:47:48 +0000 (17:47 +0300)] 
build: do not generate mech-list.h if p11-kit is not available

Compiling GnuTLS with no p11-kit installed will result in a serie of
warnings during build time because mech-list.h will be generated even if
pkcs11 tool compilation is disabled. Move mech-list.h generation to
happen only if pkcs11 is enabled, thus removing these warnings.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
7 years agobuild: pass NETTLE_LIBS together with HOGWEED_LIBS
Dmitry Eremin-Solenikov [Sat, 2 Feb 2019 14:32:01 +0000 (17:32 +0300)] 
build: pass NETTLE_LIBS together with HOGWEED_LIBS

libhogweed might depend on exact non-system-wide nettle, so let's pass
NETTLE_LIBS flags together when using HOGWEED_LIBS.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
7 years agoMerge branch 'tmp-fix-sni-error' into 'master'
Tim Rühsen [Thu, 31 Jan 2019 11:55:51 +0000 (11:55 +0000)] 
Merge branch 'tmp-fix-sni-error' into 'master'

Amend error code when SNI name is not accepted

Closes #683

See merge request gnutls/gnutls!891

7 years agoAdd GNUTLS_E_RECEIVED_DISALLOWED_NAME for illegal SNI names
Tim Rühsen [Sat, 26 Jan 2019 20:44:28 +0000 (21:44 +0100)] 
Add GNUTLS_E_RECEIVED_DISALLOWED_NAME for illegal SNI names

An illegal/disallowed SNI server name previously generated
the misleading message "An illegal parameter has been received.".

This commit changes it to
  "A disallowed SNI server name has been received.".

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'nettle-stdint' into 'master'
Tim Rühsen [Wed, 30 Jan 2019 21:32:20 +0000 (21:32 +0000)] 
Merge branch 'nettle-stdint' into 'master'

lib/nettle: replace nettle-stdint.h with just stdint.h

See merge request gnutls/gnutls!901

7 years agoMerge branch 'tmp-update-glimport-and-docs' into 'master'
Tim Rühsen [Wed, 30 Jan 2019 21:11:03 +0000 (21:11 +0000)] 
Merge branch 'tmp-update-glimport-and-docs' into 'master'

Fix 'make glimport' and update CONTRIBUTING.md

See merge request gnutls/gnutls!900

7 years agolib/nettle: replace nettle-stdint.h with just stdint.h
Dmitry Eremin-Solenikov [Wed, 30 Jan 2019 18:58:34 +0000 (21:58 +0300)] 
lib/nettle: replace nettle-stdint.h with just stdint.h

Nettle library is going to drop nettle-stdint.h. Replace this include
with with just <stdint.h>.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
7 years agoFix 'make glimport' and update CONTRIBUTING.md
Tim Rühsen [Mon, 28 Jan 2019 14:25:30 +0000 (15:25 +0100)] 
Fix 'make glimport' and update CONTRIBUTING.md

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-fix-guile-unused-var' into 'master'
Dmitry Eremin-Solenikov [Mon, 28 Jan 2019 12:46:41 +0000 (12:46 +0000)] 
Merge branch 'tmp-fix-guile-unused-var' into 'master'

Fix unused var warning in guile/src/core.c

See merge request gnutls/gnutls!895

7 years agoMerge branch 'guile' into 'master'
Tim Rühsen [Sun, 27 Jan 2019 17:23:52 +0000 (17:23 +0000)] 
Merge branch 'guile' into 'master'

build: detect previous supported guile

See merge request gnutls/gnutls!898

7 years agoMerge branch 'gitignore' into 'master'
Tim Rühsen [Sun, 27 Jan 2019 17:21:48 +0000 (17:21 +0000)] 
Merge branch 'gitignore' into 'master'

.gitignore: add test files

See merge request gnutls/gnutls!899

7 years ago.gitignore: add test files
Alon Bar-Lev [Sun, 27 Jan 2019 11:59:56 +0000 (13:59 +0200)] 
.gitignore: add test files

Signed-off-by: Alon Bar-Lev <alon.barlev@gmail.com>
7 years agobuild: detect previous supported guile
Alon Bar-Lev [Sun, 27 Jan 2019 11:17:35 +0000 (13:17 +0200)] 
build: detect previous supported guile

A recent change in the m4 macro of guile enforces latest guile:
---
AC_DEFUN([GUILE_PROGS],
 [_guile_required_version="m4_default([$1], [$GUILE_EFFECTIVE_VERSION])"
  if test -z "$_guile_required_version"; then
    _guile_required_version=2.2
  fi
---

The result:
---
checking for guile-snarf... /usr/bin/guile-snarf
checking for guild... /usr/bin/guild
checking for guile-2.2... no
checking for guile2.2... no
checking for guile-2... no
checking for guile2... no
checking for guile... /usr/bin/guile
checking for Guile version >= 2.2... configure: error: Guile 2.2 required, but 2.0.14 found
---

Probably best to specify the supported version explicitly when calling
GUILE_PROGS, to keep existing behavior calling the GUILE_PKG detects the
existing packages.

Signed-off-by: Alon Bar-Lev <alon.barlev@gmail.com>
7 years agoFix unused var warning in guile/src/core.c
Tim Rühsen [Fri, 25 Jan 2019 10:51:56 +0000 (11:51 +0100)] 
Fix unused var warning in guile/src/core.c

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoMerge branch 'tmp-fix-abi-check' into 'master'
Nikos Mavrogiannopoulos [Fri, 25 Jan 2019 17:12:22 +0000 (17:12 +0000)] 
Merge branch 'tmp-fix-abi-check' into 'master'

Fix abi-check failure

See merge request gnutls/gnutls!896

7 years agoFix abi-check failure
Tim Rühsen [Fri, 25 Jan 2019 11:26:46 +0000 (12:26 +0100)] 
Fix abi-check failure

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agoNEWS: updated gnutls_3_6_6
Nikos Mavrogiannopoulos [Fri, 25 Jan 2019 07:32:17 +0000 (08:32 +0100)] 
NEWS: updated

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agosrc/Makefile.am: remove .bak files before autogenerating
Nikos Mavrogiannopoulos [Fri, 25 Jan 2019 07:24:21 +0000 (08:24 +0100)] 
src/Makefile.am: remove .bak files before autogenerating

This avoids errors due to files pre-existing but not being
writable.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agobumped versions
Nikos Mavrogiannopoulos [Fri, 25 Jan 2019 07:18:01 +0000 (08:18 +0100)] 
bumped versions

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoMakefile.am: require guile-2.2 for release
Nikos Mavrogiannopoulos [Fri, 25 Jan 2019 07:13:26 +0000 (08:13 +0100)] 
Makefile.am: require guile-2.2 for release

That's because guile.m4 from previous releases has issues
with the latest version.

Resolves: #631

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agopriorities: when %NO_EXTENSIONS is specified disable TLS1.3
Nikos Mavrogiannopoulos [Thu, 24 Jan 2019 19:25:59 +0000 (20:25 +0100)] 
priorities: when %NO_EXTENSIONS is specified disable TLS1.3

This makes the behavior of this priority string option well-defined
even when TLS1.3 is enabled.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoMerge branch 'tmp-fix-no-extensions' into 'master'
Nikos Mavrogiannopoulos [Thu, 24 Jan 2019 19:13:52 +0000 (19:13 +0000)] 
Merge branch 'tmp-fix-no-extensions' into 'master'

The flag %NO_EXTENSIONS is disabling extension support while being functional

See merge request gnutls/gnutls!870

7 years agoMerge branch 'tmp-ametzler-certtool-manpage-formatting' into 'master'
Tim Rühsen [Thu, 24 Jan 2019 18:55:41 +0000 (18:55 +0000)] 
Merge branch 'tmp-ametzler-certtool-manpage-formatting' into 'master'

certtool.1: fix formatting

See merge request gnutls/gnutls!892

7 years agocerttool.1: fix formatting
Andreas Metzler [Thu, 24 Jan 2019 17:48:40 +0000 (18:48 +0100)] 
certtool.1: fix formatting

Apostroph at start of a line is a control character in manpages, avoid
it. Also drop wrong indent.
See https://bugs.debian.org/920215

Signed-off-by: Andreas Metzler <ametzler@bebt.de>
7 years agoMerge branch 'tmp-record-size-limit-fixes' into 'master'
Daiki Ueno [Thu, 24 Jan 2019 15:03:30 +0000 (15:03 +0000)] 
Merge branch 'tmp-record-size-limit-fixes' into 'master'

Fix record_size_limit extension handling when resuming

See merge request gnutls/gnutls!886

7 years agotlsfuzzer: update to the latest upstream for record_size_limit tests
Daiki Ueno [Wed, 23 Jan 2019 16:52:47 +0000 (17:52 +0100)] 
tlsfuzzer: update to the latest upstream for record_size_limit tests

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoMerge branch 'tmp-fix-macosx' into 'master'
Nikos Mavrogiannopoulos [Thu, 24 Jan 2019 12:24:55 +0000 (12:24 +0000)] 
Merge branch 'tmp-fix-macosx' into 'master'

.travis.yml: make macosx builds compile again

See merge request gnutls/gnutls!890

7 years agoconfigure.ac: fix substitution for libatomic
Dmitry Eremin-Solenikov [Wed, 23 Jan 2019 22:57:13 +0000 (01:57 +0300)] 
configure.ac: fix substitution for libatomic

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
7 years agoMerge branch 'tmp-fix-udp-serv' into 'master'
Dmitry Eremin-Solenikov [Wed, 23 Jan 2019 22:48:44 +0000 (22:48 +0000)] 
Merge branch 'tmp-fix-udp-serv' into 'master'

gnutls-serv: improvements in UDP server

Closes #632

See merge request gnutls/gnutls!863

7 years agoMerge branch 'tmp-priority-linear' into 'master'
Tim Rühsen [Wed, 23 Jan 2019 21:27:25 +0000 (21:27 +0000)] 
Merge branch 'tmp-priority-linear' into 'master'

set_ciphersuite_list(): Use linear approach to cleanup priorities

Closes #679

See merge request gnutls/gnutls!889

7 years ago.travis.yml: avoid installing submodules
Nikos Mavrogiannopoulos [Wed, 23 Jan 2019 19:51:11 +0000 (20:51 +0100)] 
.travis.yml: avoid installing submodules

They are not necessary for building and testing the basic
test suite.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoupdate on "Fix gnutls.pc for multiarch builds"
Nikos Mavrogiannopoulos [Wed, 23 Jan 2019 19:42:34 +0000 (20:42 +0100)] 
update on "Fix gnutls.pc for multiarch builds"

This replaces LTLIBUNISTRING with LIBUNISTRING in Makefile.am.
The former is no longer produced by configure.ac.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@gnutls.org>
7 years agoMerge branch 'master' into 'master'
Nikos Mavrogiannopoulos [Wed, 23 Jan 2019 19:14:01 +0000 (19:14 +0000)] 
Merge branch 'master' into 'master'

configure.ac: check if libatomic is needed

See merge request gnutls/gnutls!878

7 years agoMerge branch 'tmp-test-tickets' into 'master'
Nikos Mavrogiannopoulos [Wed, 23 Jan 2019 19:12:13 +0000 (19:12 +0000)] 
Merge branch 'tmp-test-tickets' into 'master'

tests: added tests for multiple ticket reception

Closes #511

See merge request gnutls/gnutls!887

7 years agoset_ciphersuite_list(): Use linear approach to cleanup priorities
Tim Rühsen [Wed, 23 Jan 2019 14:13:12 +0000 (15:13 +0100)] 
set_ciphersuite_list(): Use linear approach to cleanup priorities

Signed-off-by: Tim Rühsen <tim.ruehsen@gmx.de>
7 years agotests: check record_size_limit is reset after resumption
Daiki Ueno [Tue, 22 Jan 2019 14:47:39 +0000 (15:47 +0100)] 
tests: check record_size_limit is reset after resumption

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoconstate: don't restore max_record_recv_size from resumed data
Daiki Ueno [Sat, 19 Jan 2019 09:31:52 +0000 (10:31 +0100)] 
constate: don't restore max_record_recv_size from resumed data

Signed-off-by: Daiki Ueno <dueno@redhat.com>
7 years agoext/record_size_limit: mark it as mandatory extension
Daiki Ueno [Thu, 17 Jan 2019 16:50:49 +0000 (17:50 +0100)] 
ext/record_size_limit: mark it as mandatory extension

In a resuming session record_size_limit is always renegotiated, and
thus the server should parse the extension always.

Signed-off-by: Daiki Ueno <dueno@redhat.com>