]> git.ipfire.org Git - thirdparty/gnutls.git/log
thirdparty/gnutls.git
10 years agotests: corrected create-chain.sh to remove the ocsp_signing_key from generated certs
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 16:01:15 +0000 (18:01 +0200)] 
tests: corrected create-chain.sh to remove the ocsp_signing_key from generated certs

10 years agotests: replaced tls feature extension checks
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 16:00:03 +0000 (18:00 +0200)] 
tests: replaced tls feature extension checks

The previous checks had incorrect key purpose check on the final
(root) certificate.

10 years agoenhanced debugging messages for cert verification
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 15:31:30 +0000 (17:31 +0200)] 
enhanced debugging messages for cert verification

10 years agox509: print serial number in compact output
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 15:09:54 +0000 (17:09 +0200)] 
x509: print serial number in compact output

10 years agotests: include softhsm.h into dist files
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 14:49:42 +0000 (16:49 +0200)] 
tests: include softhsm.h into dist files

10 years agopkcs11: correctly encode the serial number when searching for certificate
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 13:27:05 +0000 (15:27 +0200)] 
pkcs11: correctly encode the serial number when searching for certificate

In gnutls_pkcs11_crt_is_known() corrected the encoding of the
serial number to TLV DER from LV DER. This is the encoding we
use when storing that number.

10 years agopkcs11: correctly account check_found_cert()
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 13:50:51 +0000 (15:50 +0200)] 
pkcs11: correctly account check_found_cert()

10 years agognutls-cli-debug: replaced draft-ietf-tls-chacha20-poly1305-04 with RFC7905
Nikos Mavrogiannopoulos [Wed, 22 Jun 2016 11:50:48 +0000 (13:50 +0200)] 
gnutls-cli-debug: replaced draft-ietf-tls-chacha20-poly1305-04 with RFC7905

10 years agognutls-cli: benchmark the memcpy performance to compare with ciphers
Nikos Mavrogiannopoulos [Mon, 20 Jun 2016 10:07:03 +0000 (12:07 +0200)] 
gnutls-cli: benchmark the memcpy performance to compare with ciphers

Also ensure that we use different memory areas for each operation
to avoid measuring better performance due to caching.

10 years agodoc update [ci skip]
Nikos Mavrogiannopoulos [Thu, 16 Jun 2016 09:53:55 +0000 (11:53 +0200)] 
doc update [ci skip]

10 years agodoc: corrected typo
Nikos Mavrogiannopoulos [Sun, 19 Jun 2016 13:48:21 +0000 (15:48 +0200)] 
doc: corrected typo

10 years agoSync with TP.
Nikos Mavrogiannopoulos [Sun, 19 Jun 2016 11:06:15 +0000 (13:06 +0200)] 
Sync with TP.

10 years agoTypo fixes (found by lintian): extention, reencode
Andreas Metzler [Sat, 18 Jun 2016 16:36:13 +0000 (18:36 +0200)] 
Typo fixes (found by lintian): extention, reencode

10 years agotests: added check for handshake packet reconstruction
Nikos Mavrogiannopoulos [Thu, 16 Jun 2016 09:51:42 +0000 (11:51 +0200)] 
tests: added check for handshake packet reconstruction

This tests whether a split handshake packet is properly reconstructed
if the parts are switched.

10 years agodtls: corrected reconstruction of handshake packets received out of order
Nikos Mavrogiannopoulos [Thu, 16 Jun 2016 09:37:30 +0000 (11:37 +0200)] 
dtls: corrected reconstruction of handshake packets received out of order

That is, when the handshake packet is split into multiple different
chunks and received out of order, make sure that reconstruction occurs
properly. Reported by Guillaume Roguez.

10 years agoCorrected the writing of serial number in PKCS#11 modules
Nikos Mavrogiannopoulos [Thu, 16 Jun 2016 08:08:34 +0000 (10:08 +0200)] 
Corrected the writing of serial number in PKCS#11 modules

That is previously the serial number was written in raw format, but
in PKCS#11 the serial number must be set encoded as integer. Report
and fix by Stanislav Zidek.

10 years agoext: ecc: replaced SUPPORTED ECC POINT FORMATS with better formatted name
Nikos Mavrogiannopoulos [Wed, 15 Jun 2016 11:50:54 +0000 (13:50 +0200)] 
ext: ecc: replaced SUPPORTED ECC POINT FORMATS with better formatted name

10 years agotests: disable SRP-base64 encode/decoded tests when SRP is disabled
Nikos Mavrogiannopoulos [Wed, 15 Jun 2016 07:21:10 +0000 (09:21 +0200)] 
tests: disable SRP-base64 encode/decoded tests when SRP is disabled

10 years ago.gitlab-ci.yml: restrict windows build checks to tests/ subdir [ci skip]
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 17:46:55 +0000 (19:46 +0200)] 
.gitlab-ci.yml: restrict windows build checks to tests/ subdir [ci skip]

That is because there is an issue with the gnulib self tests when
run under windows.

10 years agodoc update gnutls_3_5_1
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 14:43:50 +0000 (16:43 +0200)] 
doc update

10 years agodoc update
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 14:35:55 +0000 (16:35 +0200)] 
doc update

10 years agoreleased 3.5.1
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 14:32:21 +0000 (16:32 +0200)] 
released 3.5.1

10 years agotests: added missing files
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 14:31:35 +0000 (16:31 +0200)] 
tests: added missing files

10 years agoupdated auto-generated files
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 13:24:48 +0000 (15:24 +0200)] 
updated auto-generated files

10 years agotests: fixed the path of cert-tests files and added missing files in Makefile.am
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 13:23:04 +0000 (15:23 +0200)] 
tests: fixed the path of cert-tests files and added missing files in Makefile.am

10 years agomore files to ignore
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 13:17:33 +0000 (15:17 +0200)] 
more files to ignore

10 years agotests: verify the resilience of the TLSFeature handling functions on large number...
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 08:21:54 +0000 (10:21 +0200)] 
tests: verify the resilience of the TLSFeature handling functions on large number of features

10 years agotlsfeature: impose a maximum number of supported TLS features
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 08:02:56 +0000 (10:02 +0200)] 
tlsfeature: impose a maximum number of supported TLS features

This avoids many allocations and simplifies handling of the features.
The currently set maximum number of TLS features aligns with the
maximum number of supported TLS extensions.

10 years agotests: added unit test for gnutls_x509_tlsfeatures_check_crt
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 07:51:01 +0000 (09:51 +0200)] 
tests: added unit test for gnutls_x509_tlsfeatures_check_crt

10 years agoDuring PKIX chain verification check the TLSFeatures compliance
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:07:44 +0000 (16:07 +0200)] 
During PKIX chain verification check the TLSFeatures compliance

This verifies whether a chain complies with RFC7366 p.4.2.2 requirements.
That is whether the issuer's features are a superset of the certificate
under verification.

This enhances gnutls_x509_crt_get_tlsfeatures() to allow appending
of TLSFeatures, and introduces gnutls_x509_tlsfeatures_check_crt().

10 years agoverify_crt: moved all verification state into a common structure
Nikos Mavrogiannopoulos [Tue, 31 May 2016 13:10:27 +0000 (15:10 +0200)] 
verify_crt: moved all verification state into a common structure

This allows for easier extension of state.

10 years agotests: added chain verification with TLS features
Nikos Mavrogiannopoulos [Tue, 31 May 2016 07:28:28 +0000 (09:28 +0200)] 
tests: added chain verification with TLS features

That adds checks for the RFC7633 requirements for intermediate
and CA certificates (p. 4.2.2).

10 years agobumped version
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 12:37:49 +0000 (14:37 +0200)] 
bumped version

10 years agodoc update
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 12:37:12 +0000 (14:37 +0200)] 
doc update

10 years agodoc update
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 12:01:56 +0000 (14:01 +0200)] 
doc update

10 years agotests: verify the operation of honor_crq_ext template option
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 11:49:28 +0000 (13:49 +0200)] 
tests: verify the operation of honor_crq_ext template option

10 years agotests: common.sh will export the required TZ for datefudge tests
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 11:41:24 +0000 (13:41 +0200)] 
tests: common.sh will export the required TZ for datefudge tests

10 years agotools: avoid using deprecated types
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 11:24:03 +0000 (13:24 +0200)] 
tools: avoid using deprecated types

10 years agocerttool: allow copying specific certificate request extensions to certificate
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 11:23:03 +0000 (13:23 +0200)] 
certtool: allow copying specific certificate request extensions to certificate

This introduces the honor_crq_extension multi-line template option.

10 years agotests: added check on gnutls_x509_crt_set_crq_extension_by_oid()
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 11:13:32 +0000 (13:13 +0200)] 
tests: added check on gnutls_x509_crt_set_crq_extension_by_oid()

10 years agoAdded gnutls_x509_crt_set_crq_extension_by_oid()
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 09:19:50 +0000 (11:19 +0200)] 
Added gnutls_x509_crt_set_crq_extension_by_oid()

This allows copying specific OIDs from a certificate
request to the certificate.

10 years agotests: moved check for datefudge in scripts/common.sh
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 07:59:20 +0000 (09:59 +0200)] 
tests: moved check for datefudge in scripts/common.sh

10 years agotests: cert-tests: moved all data files in separate subdir
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 07:55:04 +0000 (09:55 +0200)] 
tests: cert-tests: moved all data files in separate subdir

10 years agotests: cert-tests: moved templates into subdir
Nikos Mavrogiannopoulos [Tue, 14 Jun 2016 07:36:29 +0000 (09:36 +0200)] 
tests: cert-tests: moved templates into subdir

10 years agotests: test trailing comma in system priorities
Daniel P. Berrange [Fri, 10 Jun 2016 12:41:22 +0000 (13:41 +0100)] 
tests: test trailing comma in system priorities

Add tests which verify behaviour when the list of system
priorities has a trailing ','. Avoid crash in test suite
if the test unexpectedly succeeds when expected_str is
NULL.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
10 years agotests: added check of DTLS rehandshake for upgrade
Nikos Mavrogiannopoulos [Fri, 10 Jun 2016 11:56:23 +0000 (13:56 +0200)] 
tests: added check of DTLS rehandshake for upgrade

That is check whether anon -> cert renegotiation works.

10 years agotests: added check of DTLS rehandshake when using PKIX certs
Nikos Mavrogiannopoulos [Fri, 10 Jun 2016 11:43:36 +0000 (13:43 +0200)] 
tests: added check of DTLS rehandshake when using PKIX certs

This complements the existing DTLS rehandshake test using anonymous
ciphersuites.

10 years agotests: document some details in system-prio-file [ci skip]
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 16:55:02 +0000 (18:55 +0200)] 
tests: document some details in system-prio-file [ci skip]

10 years agodoc: mention the usage of the _gnutls_resolve_priorities function in testsuite
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:40:49 +0000 (17:40 +0200)] 
doc: mention the usage of the _gnutls_resolve_priorities function in testsuite

10 years agodoc: mention the fallback keyword support in manual
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:38:29 +0000 (17:38 +0200)] 
doc: mention the fallback keyword support in manual

10 years agotests: added checks for system priority file fallback mechanism
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:36:12 +0000 (17:36 +0200)] 
tests: added checks for system priority file fallback mechanism

10 years agognutls_priority_init: multiple @KEYWORD lookups with fallback
Daniel P. Berrange [Fri, 3 Jun 2016 15:57:02 +0000 (16:57 +0100)] 
gnutls_priority_init: multiple @KEYWORD lookups with fallback

The support for using "@KEYWORD" as a priority string is very
useful to separate selection of priorities from application
specific code or config files. It is, however, not general
enough to fully serve all reasonable use cases.

For example, consider an application sets

  gnutls_priority_set_direct(session, "@SYSTEM", NULL);

The system administrator can modify the global priorities
file to change what "@SYSTEM" resolves to for all apps using
GNUTLS. As soon as one application wishes to have a slightly
different configuration from others on the host, you have to
go back and start modifying application specific configuration
files once more. This is bad for the system administrator as
it means there's no longer one single place where they can
see the priority configuration for all apps.

They may try to get around this problem by configuring the
app to use a different keyword, instead of a full priority
string, eg "@LIBVIRT". So the global priorities file can
now define entries for both "SYSTEM" and "LIBVIRT". This
has still placed a burden on the administrator change the
config in two places - both libvirt config files and the
global priorities file.

What is more desirable is if applications were able to provide
a list of keywords that would be tried in order, picking the
first that existed. For example, libvirt could be written to
request the following by default

  gnutls_priority_set_direct(session, "@LIBVIRT,SYSTEM", NULL);

With this, gnutls would first try to find the "LIBVIRT" keyword
in the global configuration file, and if that is not present,
then it would fallback to trying to find the "SYSTEM" keyword.

This provides nice "out of the box" behaviour for system
administrators, whereby the app would be using "SYSTEM" initially
and if the admin wishes to give the app a custom configuration,
they can simply modify the global priorities file to add in the
application specific keyword "LIBVIRT". There is never a need
for the sysadmin to modify any application specific configuration
files any more. It is exclusively controlled in one place via the
global priorities file.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
10 years agotests: enhanced system priority file testing
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:20:22 +0000 (17:20 +0200)] 
tests: enhanced system priority file testing

This checks whether appending to system priority options work.

10 years agodoc update
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:21:57 +0000 (17:21 +0200)] 
doc update

10 years agodoc: remove all references to openpgp auth example
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 15:01:06 +0000 (17:01 +0200)] 
doc: remove all references to openpgp auth example

10 years agodoc update
Nikos Mavrogiannopoulos [Thu, 9 Jun 2016 14:48:14 +0000 (16:48 +0200)] 
doc update

10 years ago_gnutls_resolve_priorities: always try to re-read sys priority file
Daniel P. Berrange [Fri, 3 Jun 2016 13:59:11 +0000 (14:59 +0100)] 
_gnutls_resolve_priorities: always try to re-read sys priority file

Previously if the system priority file was edited, that would
take effect on the very next TLS session an application created.

As of:

  commit 006b89d4464ae1bb6d545ea5716998654124df45
  Author: Nikos Mavrogiannopoulos <nmav@redhat.com>
  Date:   Fri Apr 1 10:46:12 2016 +0200

    priorities: preload the system priorities on library loading time

It is required to restart every application after changing the
system priority file to get changes to take effect.

Further, for applications running in a chroot, it will no longer
honour a system priority file that may exist inside the chroot,
always using the originally cached data from outside the chroot.

This patch changes the caching so that we always try to reload
the cache of system priorities. A mtime check is used to avoid
actually re-reading the file unless its content has obviously
changed. If the file no longer exists, the cache will not be
invalidated. This ensures that the current priority file is
always honoured, whether inside a chroot or not, while at the
same time allowing apps to work in a chroot when no system
priority file is present.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
10 years agodoc: remove references to GNUTLS_KEYLOGFILE
Nikos Mavrogiannopoulos [Wed, 8 Jun 2016 17:38:07 +0000 (19:38 +0200)] 
doc: remove references to GNUTLS_KEYLOGFILE

10 years agotests: renamed rehandshake checks for clarity
Nikos Mavrogiannopoulos [Mon, 6 Jun 2016 17:33:34 +0000 (19:33 +0200)] 
tests: renamed rehandshake checks for clarity

10 years agodoc update
Nikos Mavrogiannopoulos [Mon, 6 Jun 2016 16:47:28 +0000 (18:47 +0200)] 
doc update

10 years agokeylogfile: only consider the SSLKEYLOGFILE variable
Nikos Mavrogiannopoulos [Mon, 6 Jun 2016 16:43:12 +0000 (18:43 +0200)] 
keylogfile: only consider the SSLKEYLOGFILE variable

In addition do not check the environment in the constructor
but instead use static variables to save the key file name.
The GNUTLS_KEYLOGFILE environment variable is no longer used
since there is no reason to have a separate one.

10 years agolib: eliminated the use of deprecated variables
Nikos Mavrogiannopoulos [Mon, 6 Jun 2016 16:39:18 +0000 (18:39 +0200)] 
lib: eliminated the use of deprecated variables

10 years agodoc: removed OpenPGP examples
Nikos Mavrogiannopoulos [Sat, 4 Jun 2016 09:42:21 +0000 (11:42 +0200)] 
doc: removed OpenPGP examples

Relates #102

10 years agopkcs12: corrected return type of gnutls_pkcs12_bag_get_type()
Nikos Mavrogiannopoulos [Sat, 4 Jun 2016 09:36:34 +0000 (11:36 +0200)] 
pkcs12: corrected return type of gnutls_pkcs12_bag_get_type()

10 years agotests: move pkcs11-cert-import-url4-exts with the other pkcs11 tests
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 15:07:28 +0000 (17:07 +0200)] 
tests: move pkcs11-cert-import-url4-exts with the other pkcs11 tests

This prevents a build failure in windows.

10 years agodoc clarify the version since when GNUTLS_PKCS11_OBJ_FLAG_OVERWRITE_TRUSTMOD_EXT...
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 13:01:59 +0000 (15:01 +0200)] 
doc clarify the version since when GNUTLS_PKCS11_OBJ_FLAG_OVERWRITE_TRUSTMOD_EXT is accepted

10 years agotests: corrected typo in crl-test
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 12:53:04 +0000 (14:53 +0200)] 
tests: corrected typo in crl-test

10 years agotests: check gnutls_pkcs11_obj_list_import_url4() with GNUTLS_PKCS11_OBJ_FLAG_OVERWRI...
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 12:24:32 +0000 (14:24 +0200)] 
tests: check gnutls_pkcs11_obj_list_import_url4() with GNUTLS_PKCS11_OBJ_FLAG_OVERWRITE_TRUSTMOD_EXT

10 years agognutls_pkcs11_obj_list_import_url4: accepts the GNUTLS_PKCS11_OBJ_FLAG_OVERWRITE_TRUS...
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 12:50:45 +0000 (14:50 +0200)] 
gnutls_pkcs11_obj_list_import_url4: accepts the GNUTLS_PKCS11_OBJ_FLAG_OVERWRITE_TRUSTMOD_EXT flag

10 years agognutls_pkcs11_obj_list_import_url3: rewritten to use gnutls_pkcs11_obj_list_import_url4
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 12:41:25 +0000 (14:41 +0200)] 
gnutls_pkcs11_obj_list_import_url3: rewritten to use gnutls_pkcs11_obj_list_import_url4

10 years agopkcs11: use ctx as variable name for ck_object_handle_t for clarity
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 12:29:45 +0000 (14:29 +0200)] 
pkcs11: use ctx as variable name for ck_object_handle_t for clarity

10 years agopkcs11: doc update
Nikos Mavrogiannopoulos [Fri, 3 Jun 2016 11:47:44 +0000 (13:47 +0200)] 
pkcs11: doc update

10 years agodoc update [ci skip]
Nikos Mavrogiannopoulos [Thu, 2 Jun 2016 12:37:11 +0000 (14:37 +0200)] 
doc update [ci skip]

10 years ago_gnutls_check_key_purpose: in CA certificates treat the SGC key purpose as GNUTLS_KP_...
Nikos Mavrogiannopoulos [Thu, 2 Jun 2016 12:28:52 +0000 (14:28 +0200)] 
_gnutls_check_key_purpose: in CA certificates treat the SGC key purpose as GNUTLS_KP_TLS_WWW_SERVER

This is a hack for certain very old CA certificates lurking around which instead
of having the GNUTLS_KP_TLS_WWW_SERVER have some old OIDs for that purpose. Consider
these OIDs equivalent to GNUTLS_KP_TLS_WWW_SERVER in marked as CA certificates.

10 years agognutls-cli: --save-ocsp will work even if verification fails
Nikos Mavrogiannopoulos [Thu, 2 Jun 2016 07:22:26 +0000 (09:22 +0200)] 
gnutls-cli: --save-ocsp will work even if verification fails

That is, allow saving the response even if the OCSP response caused
a verification error. That way the response can be examined for possible
issues.

10 years agoocsp: attempt harder to figure an OCSP staple issuer
Nikos Mavrogiannopoulos [Thu, 2 Jun 2016 07:20:13 +0000 (09:20 +0200)] 
ocsp: attempt harder to figure an OCSP staple issuer

That is, check initially against the trust list set on the credentials,
and if verification is not possible attempt with all certificates in
the chain as possible issuers. The reason of this enhancement is the
few servers have an OCSP response signed not by their direct CA but
rather by one of the higher level CAs.

10 years agotests: added comprehensive OCSP test suite with MUST-staple PKIX extension
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 12:10:59 +0000 (14:10 +0200)] 
tests: added comprehensive OCSP test suite with MUST-staple PKIX extension

This includes the tests:
 - Server with valid certificate - no staple
 - Server with valid certificate - valid staple
 - Server with valid certificate - invalid staple
 - Server with valid certificate - unrelated cert staple
 - Server with valid certificate - expired staple
 - Server with valid certificate - old staple

10 years agotests: utils: added c_print()
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 08:09:54 +0000 (10:09 +0200)] 
tests: utils: added c_print()

10 years agoext: status_request: added more descriptive name
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 08:33:15 +0000 (10:33 +0200)] 
ext: status_request: added more descriptive name

10 years agoocsp: fail certificate verification on expired or too old revocation data info
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 12:15:49 +0000 (14:15 +0200)] 
ocsp: fail certificate verification on expired or too old revocation data info

10 years agoocsp: Introduced GNUTLS_CERT_INVALID_OCSP_STATUS
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 12:13:52 +0000 (14:13 +0200)] 
ocsp: Introduced GNUTLS_CERT_INVALID_OCSP_STATUS

This verification status flag indicates an OCSP status response
being stapled but it being invalid for some reason (e.g., unable
to parse or doesn't contain the expected certificate).

10 years agodoc: improved OCSP description and mention RFC7633
Nikos Mavrogiannopoulos [Wed, 1 Jun 2016 12:38:32 +0000 (14:38 +0200)] 
doc: improved OCSP description and mention RFC7633

10 years agotests: added basic check for gnutls_url_is_supported
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:52:06 +0000 (16:52 +0200)] 
tests: added basic check for gnutls_url_is_supported

10 years agognutls_url_is_supported: type changed to unsigned
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:47:34 +0000 (16:47 +0200)] 
gnutls_url_is_supported: type changed to unsigned

In addition function documentation was updated.

10 years agodoc update
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:42:22 +0000 (16:42 +0200)] 
doc update

10 years agopubkey_to_bits: return type was changed to unsigned
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:37:34 +0000 (16:37 +0200)] 
pubkey_to_bits: return type was changed to unsigned

This function did not return signed data, so the "int" return
type was confusing.

10 years agocrypto-selftests: removed unneeded cast
Nikos Mavrogiannopoulos [Tue, 31 May 2016 14:16:38 +0000 (16:16 +0200)] 
crypto-selftests: removed unneeded cast

10 years agoseveral sign-related API changes
Nikos Mavrogiannopoulos [Tue, 31 May 2016 11:49:13 +0000 (13:49 +0200)] 
several sign-related API changes

This replaces the usage of "int" in functions which could only
have accepted an "unsigned" value. Also functions which return
unsigned values are explicitly tagged as such. The ABI remains
the same with these changes.

This allows easier catching of sign/unsigned related errors from
the calling applications.

10 years agox509: simplified _gnutls_x509_get_tbs()
Nikos Mavrogiannopoulos [Tue, 31 May 2016 09:40:00 +0000 (11:40 +0200)] 
x509: simplified _gnutls_x509_get_tbs()

10 years agox509: replace the bool type with the unsigned type
Nikos Mavrogiannopoulos [Tue, 31 May 2016 08:08:14 +0000 (10:08 +0200)] 
x509: replace the bool type with the unsigned type

This allows to rely on gcc warnings for improper checks and
conversions. Unfortunately gcc does warn on invalid checks for
the bool type (e.g., b<0).

10 years agoconfigure: enable the type-limits gcc warnings
Nikos Mavrogiannopoulos [Tue, 31 May 2016 08:00:18 +0000 (10:00 +0200)] 
configure: enable the type-limits gcc warnings

In addition remove the unsafe-loop-optimizations warning
as they were not helpful.

10 years agocerttool: doc update
Nikos Mavrogiannopoulos [Tue, 31 May 2016 06:59:47 +0000 (08:59 +0200)] 
certtool: doc update

10 years agoDCO: added Tim Kosse [ci skip]
Nikos Mavrogiannopoulos [Mon, 30 May 2016 20:28:23 +0000 (22:28 +0200)] 
DCO: added Tim Kosse [ci skip]

10 years agotools: modify canonicalize_host to not depend on in6_addr
Nikos Mavrogiannopoulos [Mon, 30 May 2016 13:19:15 +0000 (15:19 +0200)] 
tools: modify canonicalize_host to not depend on in6_addr

10 years agotests: added unit tests for gnutls_x509_tlsfeatures_t handling funcs
Nikos Mavrogiannopoulos [Mon, 30 May 2016 13:17:06 +0000 (15:17 +0200)] 
tests: added unit tests for gnutls_x509_tlsfeatures_t handling funcs

This includes DER import/export as well as feature appending.

10 years agotests: tlsfeature-test will ignore the 'Algorithm Security Level' line in comparisons
Nikos Mavrogiannopoulos [Mon, 30 May 2016 12:19:20 +0000 (14:19 +0200)] 
tests: tlsfeature-test will ignore the 'Algorithm Security Level' line in comparisons

That is to allow depending on the certificate output validation without
relying on "moving" parameters such as the Algorithm Security Level.

10 years agotests: verify whether the TLSFeatures extension is copied
Nikos Mavrogiannopoulos [Mon, 30 May 2016 11:47:14 +0000 (13:47 +0200)] 
tests: verify whether the TLSFeatures extension is copied

Verify whether the TLSFeatures extension is copied from the certificate
request to the generated certificate.

10 years agodoc: updated since version of tlsfeature functionality and documented new functions
Nikos Mavrogiannopoulos [Mon, 30 May 2016 11:34:17 +0000 (13:34 +0200)] 
doc: updated since version of tlsfeature functionality and documented new functions

10 years agotests: add testcase to check for missing status request
Tim Kosse [Fri, 15 Jan 2016 13:45:03 +0000 (14:45 +0100)] 
tests: add testcase to check for missing status request

That is verify whether the OCSP MUST-staple extension, as can be
deduced from RFC7633, is accounted during handshake.

10 years agoReset extensions_sent_size only at start of handshake
Tim Kosse [Fri, 15 Jan 2016 13:45:42 +0000 (14:45 +0100)] 
Reset extensions_sent_size only at start of handshake

That is, do not reset it when completing it so that we can
use the negotiated extensions even after the handshake is
complete.

Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>