Standardize file:line messages
Reduce buff size; move to end of STANZA
Add some Title entries (with blank line after)
Add Title to some BN test files.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3608)
Code was added in commit b3c31a65 that overwrote the last ex_data value
using CRYPTO_dup_ex_data() causing a memory leak, and potentially
confusing the ex_data dup() callback.
In ssl_session_dup(), fix error handling (properly reference and up-ref
shared data) and new-up the ex_data before calling CRYPTO_dup_ex_data();
all other structures that dup ex_data have the destination ex_data new'd
before the dup.
Fix up some of the ex_data documentation.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3323)
Add Ed25519 certificate verify test using certificate from
draft-ietf-curdle-pkix-04 and custom generated root certificate.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3503)
The version number 3 means version 4, while 2 means version 3. Since this is the v3nametest, version 3 should be used.
CLA: Trivial
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3577)
Recently introduced TEST_* macros print variables' symbolic names.
In order to make error output more readable rename some variables.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Commit bd990e2535 changed our handling of alerts. Some of the BoringSSl
tests were expecting specific errors to be created if bad alerts were sent.
Those errors have now changed as a result of that commit, so the BoringSSL
test config needs to be updated to match.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3549)
Excess bytes, when one block is longer than the other, are not explicitly
highlighted.
The NULL / zero length block output has been cleaned up.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3515)
Fix the small arena test to just check for the symptom of the infinite
loop (i.e. initialized set on failure), rather than the actual infinite
loop. This avoids some valgrind errors.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3512)
Switch to TAP::Harness inadvertently masked test failures.
Test::Harness::runtests was terminating with non-zero exit code in case
of failure[s], while TAP::Harness apparently holds caller responsible
for doing so.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Remove assertion when mmap() fails.
Only run the 1<<31 limit test on Linux
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3455)
Import test data from sha1test.c, sha256t.c and sha512t.c which is
from RFC6234 section 8.5
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3451)
Allow multiple "Input" lines to call the update function multiple times.
Add "Ncopy" keyword to copy the input buffer. So for example:
Input = "a"
Ncopy = 1024
Will create a buffer consisting of 1024 "a" characters.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3451)
An alert message is 2 bytes long. In theory it is permissible in SSLv3 -
TLSv1.2 to fragment such alerts across multiple records (some of which
could be empty). In practice it make no sense to send an empty alert
record, or to fragment one. TLSv1.3 prohibts this altogether and other
libraries (BoringSSL, NSS) do not support this at all. Supporting it adds
significant complexity to the record layer, and its removal is unlikely
to cause inter-operability issues.
The DTLS code for this never worked anyway and it is not supported at a
protocol level for DTLS. Similarly fragmented DTLS handshake records only
work at a protocol level where at least the handshake message header
exists within the record. DTLS code existed for trying to handle fragmented
handshake records smaller than this size. This code didn't work either so
has also been removed.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3476)
Allow multiple file arguments.
Split bntests.txt into separate files.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3471)
Issue 1:
sh.bittable_size is a size_t but i is and int, which can result in
freelist == -1 if sh.bittable_size exceeds an int.
This seems to result in an OPENSSL_assert due to invalid allocation
size, so maybe that is "ok."
Worse, if sh.bittable_size is exactly 1<<31, then this becomes an
infinite loop (because 1<<31 is a negative int, so it can be shifted
right forever and sticks at -1).
Issue 2:
CRYPTO_secure_malloc_init() sets secure_mem_initialized=1 even when
sh_init() returns 0.
If sh_init() fails, we end up with secure_mem_initialized=1 but
sh.minsize=0. If you then call secure_malloc(), which then calls,
sh_malloc(), this then enters an infite loop since 0 << anything will
never be larger than size.
Issue 3:
That same sh_malloc loop will loop forever for a size greater
than size_t/2 because i will proceed (assuming sh.minsize=16):
i=16, 32, 64, ..., size_t/8, size_t/4, size_t/2, 0, 0, 0, 0, ....
This sequence will never be larger than "size".
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3449)
In some cases, testutil outputs subtests like this:
1..6 # Subtest: progname
The standard set by Test::More (because there really is no actual
standard yet) gives this display:
# Subtest: progname
1..6
Until the standard is actually agreed upon, let's do it like
Test::More.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3446)
- Mostly missing fall thru comments
- And uninitialized value used in sslapitest.c
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3440)
Add file/line# to test error message.
Also remove expected/got fields since TEST structure prints them.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3324)
Also, allow multiple files on commandline (for future splitup of
evptests.txt)
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3324)
TAP:Harness came along in perl 5.10.1, and since we claim to support
perl 5.10.0 in configuration and testing, we can only load it
conditionally.
The main reason to use TAP::Harness rather than Test::Harness is its
capability to merge stdout and stderr output from the test recipes,
which Test::Harness can't. The merge gives much more comprehensible
output when testing verbosely.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3424)
This includes support for:
- comparisions between pairs of BIGNUMs
- comparisions between BIGNUMs and zero
- equality comparison between BIGNUMs and one
- equality comparisons between BIGNUMs and constants
- parity checks for BIGNUMs
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3405)
Verify that we fail if we receive an HRR but no change will result in
ClientHello2.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3414)
Test that sending a non NULL compression method fails in TLSv1.3 as well
as other similar tests.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3410)
Get some trivial test coverage that this flag does what it claims to.
[extended tests]
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1700)
Check we send supported_groups in EE if there is a group we prefer instead
of the one sent in the key_share.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3395)
The conditions to skip these recipes entirely don't show in a
non-verbose test harness output. We prefer to know, so use skip_all,
as it is a little bit more verbose.
[extended tests]
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3398)
Some refactoring done as well.
The prime_field_tests() function needs splitting and refactoring still.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3340)
Fix warning/bug in rc5test
Remove useless/warning-only test from dsatest.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3387)
Format the test failure output more nicely.
More vertical space is used to make things a little clearer. Tests are expected
to pass so this doesn't impact the normal case.
Strings and memory comparisons highlight differences.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3357)
Because we now have TAP output for every mod_exp round, there's no
more need to mark the round with outputting a period.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3380)
The previous commit fixed a bug which occurs when serverinfo is loaded
from memory (not from a file). This adds a test for loading serverinfo
from memory.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3382)
Add padding callback for application control
Standard block_size callback
Documentation and tests included
Configuration file/s_client/s_srver option
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3130)
ASN1_GENERALIZEDTIME and ASN1_UTCTIME may be specified using offsets,
even though that's not supported within certificates.
To convert the offset time back to GMT, the offsets are supposed to be
subtracted, not added. e.g. 1759-0500 == 2359+0100 == 2259Z.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2654)
It occurs when memory compares are made that are larger
than the on stack temporary buffers (either malloced or supplied).
Rework the test test so it doesn't use a macro with a branch.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3155)
gcc's -Wextra pulls in -Wold-style-declaration, which triggers when a
declaration has a storage-class specifier as a non-initial qualifier.
The ISO C formal grammar requires the storage-class to be the first
component of the declaration, if present.
Seeint as the register storage-class specifier does not really have any effect
anymore with modern compilers, remove it entirely while we're here, instead of
fixing up the order.
Interestingly, the gcc devteam warnings do not pull in -Wextra, though
the clang ones do.
[extended tests]
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3239)
Also added a internal error printing callback to be used both with
ERR_print_errors_cb() and with CRYPTO_mem_leaks_cb
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)
These functions aren't meant to be used directly by the test programs,
reflect that by making the declarations a little harder to reach, but
still available enough if there's a need to override them.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3345)
Enforcement of an SNI extension in the initial ClientHello is becoming
increasingly common (e.g. see GitHub issue #2580). This commit changes
s_client so that it adds SNI be default, unless explicitly told not to via
the new "-noservername" option.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2614)
Also converted most of ssltestlib but left the packet_dump output
as-is (for now).
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3257)
Updated due to test framework changes
Updates after code review
Missed some checks
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3269)
The previous commits added sanity checks for where the max enabled protocol
version does not have any configured ciphersuites. We should check that we
fail in those circumstances.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)
Ensure that there are ciphersuites enabled for the maximum supported
version we will accept in a ClientHello.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)
Ensure that there are ciphersuites enabled for the maximum supported
version we are claiming in the ClientHello.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3316)
Original rationale behind using write in testutil was to accommodate
no-stdio builds. But is there evidence that no-stdio users would have
write or pre-defined meaning for file descriptors 1 and 2? Correct
answer is to provide way for no-stdio users who want to exercise
tests to plug in own BIO, not to make assumption that they have write.
And since we don't have to make such assumption, we can as well go
for simplest that works with standard library as specified by C
language standard.
Reviewed-by: Richard Levitte <levitte@openssl.org>
This includes better signals of skips and subtests according to TAP 12,
and flushing stdout and stderr at the end of every test function to
make sure we get the output in good order.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3309)
The CA names should be printed according to user's decision
print_name instead of set of BIO_printf
dump_cert_text instead of set of BIO_printf
Testing cyrillic output of X509_CRL_print_ex
Write and use X509_CRL_print_ex
Reduce usage of X509_NAME_online
Using X509_REQ_print_ex instead of X509_REQ_print
Fix nameopt processing.
Make dump_cert_text nameopt-friendly
Move nameopt getter/setter to apps/apps.c
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3262)
With the perl test framework comes the output format TAP
(Test Anything Protocol, see http://testanything.org/) with
extra extension for subtests. This change extends that same
output format to any test program using testutil.
In this implementation, each test program is seen as a full test that
can be used as a subtest. The perl framework passes on the subtest
level to the test programs with the environment variable
HARNESS_OSSL_LEVEL. Furthermore, and series of tests added with
ADD_ALL_TESTS is regarded as another subtest level.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3296)
It's now built as a static library, and greatly simplified for test
programs, which no longer need to include test_main_custom.h or
test_main.h and link with the corresponding object files.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3243)
This includes reworked reworked tests to do both encrypt and decrypt,
and a few more ciphers added.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3197)
Previously, BN_is_prime_fasttest_ex, when doing trial-division, would
check whether the candidate is a multiple of a number of small primes
and, if so, reject it. However, three is a multiple of three yet is
still a prime number.
This change accepts small primes as prime when doing trial-division.
Reviewed-by: Kurt Roeckx <kurt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3264)
X509_STORE_add_cert and X509_STORE_add_crl are changed to return
success if the object to be added was already found in the store, rather
than returning an error.
Raise errors if empty or malformed files are read when loading certificates
and CRLs.
Remove NULL checks and allow a segv to occur.
Add error handing for all calls to X509_STORE_add_c{ert|tl}
Refactor these two routines into one.
Bring the unit test for duplicate certificates up to date using the test
framework.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2830)
Updated after code review, and fix indenting
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3175)
Note that these guards are still needed around local header files that
declare linkable symbols.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3259)
The internals test programs access header files that aren't guarded by
the public __DECC_INCLUDE_PROLOGUE.H and __DECC_INCLUDE_EPILOGUE.H files,
and therefore have no idea what the naming convention is. Therefore, we
need to specify that explicitely in the internals test programs, since
they aren't built with the same naming convention as the library they
belong with.
Reviewed-by: Tim Hudson <tjh@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3247)
ct_test,evp_extra_test,wpackettest,packettest
Add strncmp TEST wrappers
And make some style/consistency fixes to ct_test
Silence travis; gcc bug?
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3234)
randtest, cipher_overhead_test, bioprintest, constant_time_test
Move test_bioprint to 04 group
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3228)
If EC support is enabled we should catch also EC_R_UNKNOWN_GROUP as an hint to
an unsupported algorithm/curve (e.g. if binary EC support is disabled).
Before this commit the issue arise for example if binary EC keys are added in
evptests.txt, and the test is run when EC is enabled but EC2m is disabled.
E.g. adding these lines to evptests.txt would reproduce the issue:
~~~
PrivateKey=KAS-ECC-CDH_K-163_C0
-----BEGIN PRIVATE KEY-----
MGMCAQAwEAYHKoZIzj0CAQYFK4EEAAEETDBKAgEBBBUAZlO2B3OY+tx79eYBWBcB
SMPcRSehLgMsAAQHH4sod9YCfZwa3kJE8t6hJpLvI9UFwV7ndiIccrhLNHzjg/OA
Z7icPpo=
-----END PRIVATE KEY-----
PublicKey=KAS-ECC-CDH_K-163_C0-PUBLIC
-----BEGIN PUBLIC KEY-----
MEAwEAYHKoZIzj0CAQYFK4EEAAEDLAAEBx+LKHfWAn2cGt5CRPLeoSaS7yPVBcFe
53YiHHK4SzR844PzgGe4nD6a
-----END PUBLIC KEY-----
PublicKey=KAS-ECC-CDH_K-163_C0-Peer-PUBLIC
-----BEGIN PUBLIC KEY-----
MEAwEAYHKoZIzj0CAQYFK4EEAAEDLAAEBXQjbxQoxDITCUZ4Ols6q7bCfqXWB5CM
JRuNoCHLrCgfEj969PrFs9u4
-----END PUBLIC KEY-----
Derive=KAS-ECC-CDH_K-163_C0
PeerKey=KAS-ECC-CDH_K-163_C0-Peer-PUBLIC
Ctrl=ecdh_cofactor_mode:1
SharedSecret=04325bff38f1b0c83c27f554a6c972a80f14bc23bc
~~~
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3226)
When compiling without EC support the test fails abruptly reading some keys.
Some keys merged in commit db04055 start with
------BEGIN EC PRIVATE KEY-----
this format is not supported without EC support.
This commit reformat those keys with the generic format. After this change the
test simply skips the unsupported EC keys when EC is disabled, without parsing
errors.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3223)
All tests from ecdhtest.c have been ported to evptests.txt
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3219)
If $_ is not private, it can wipe caller's one, which proved to be
problematic...
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Add test case that checks some of them.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3208)
The only SSL tests prior to this tested using certificates with no
embedded Signed Certificate Timestamps (SCTs), which meant they couldn't
confirm whether Certificate Transparency checks in "strict" mode were
working.
These tests reveal a bug in the validation of SCT timestamps, which is
fixed by the next commit.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3138)
95-test_external_boringssl.t had a specialised run() variant to prefix
the command output so it wouldn't disturb Test::Harness. This
functionality if now moved to the run() command, using the added
option 'prefix' that can be set to the string to prefix the output
with.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3201)
Because stdout is usually buffered and stderr isn't, error output
might get printed in one bunch and all the lines saying which test
failed all in one bunch, making it difficult to see exactly what error
output belongs to what test. Flushing stdout makes sure the runner
output is displayed together with the corresponding error output.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3198)
information.
The framework will display the non-matching memory.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3156)
We're already checking that custom DER decodes to expected values (or
fails to do so), but we didn't check if values encode back to expected
DER.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3174)
It isn't easy to use the test framework since it turns memory debugging
on as well and the CRYPTO_mem_leaks_fp function cannot be called twice.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3169)
RFC 7301 mandates that the server SHALL respond with a fatal
"no_application_protocol" alert when there is no overlap between
the client's supplied list and the server's list of supported protocols.
In commit 062178678f we changed from
ignoring non-success returns from the supplied alpn_select_cb() to
treating such non-success returns as indicative of non-overlap and
sending the fatal alert.
In effect, this is using the presence of an alpn_select_cb() as a proxy
to attempt to determine whether the application has configured a list
of supported protocols. However, there may be cases in which an
application's architecture leads it to supply an alpn_select_cb() but
have that callback be configured to take no action on connections that
do not have ALPN configured; returning SSL_TLSEXT_ERR_NOACK from
the callback would be the natural way to do so. Unfortunately, the
aforementioned behavior change also treated SSL_TLSEXT_ERR_NOACK as
indicative of no overlap and terminated the connection; this change
supplies special handling for SSL_TLSEXT_ERR_NOACK returns from the
callback. In effect, it provides a way for a callback to obtain the
behavior that would have occurred if no callback was registered at
all, which was not possible prior to this change.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2570)
Don't compile code that still uses LONG when it's deprecated
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3126)
It's sheer luck that this was used for the first field only which also
has the same type in all data structures, so the offsets were never wrong
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3127)
Make sure the server can write normal data after earlier writing early data.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3091)
Commit 9b5c865df introduced a synthetic delay between arrival of EoED and
CF. We actually want to delay the arrival of CF even further to demonstrate
that we can write early data even when "in init".
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3091)
This test doesn't actually fail completely, but there's no real
pattern to distinguish which data files should be omitted when no-ec2m
is configured and which should not.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3103)
'j' is specified as modifier for "greatest-width integer type", which in
practice means 64 bits on both 32- and 64-bit platforms. Since we rely
on __attribute__((__format__(__printf__,...))) to sanitize BIO_print
format, we can use it to denote [u]int64_t-s in platform-neutral manner.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3083)
This tests the bug fixed in the previous commit. We introduce a synthetic
delay between the server receiving EoED and CF and check that we can still
send early data.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3089)
We want to make sure that if we if are using SSL_MODE_AUTO_RETRY then
if SSL_read_early_data() hits EndOfEarlyData then it doesn't auto retry
and end up with normal data. The same issue could occur with read_ahead
which is what we use in this test.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3077)
Fix some comments too
[skip ci]
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3069)
Remove unnecessary include of apps.h. Tests shouldn't take a
dependency on apps. In this case, there is no dependency, the include
is unnecessary.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Also, be less silent when installing, so possible errors are shown.
[extended tests]
Fixes#3005
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3007)
When using run() with capture => 1, there was no way to find out if
the command was successful or not. This change adds a statusvar
option, that must refer to a scalar variable, for example:
my $status = undef;
my @line = run(["whatever"], capture => 1, statusvar => \$status);
$status will be 1 if the command "whatever" was successful, 0
otherwise.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3004)
The code to do this incorrectly assumed that the protocol version
could be used as a valid cipher suite for the 'openssl cipher'
command. While this is true in some cases, that isn't something to be
trusted. Replace that assumption with code that takes the full
'openssl ciphers' command output and parses it to find the ciphers we
look for.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2956)
Fix a strict aliasing issue in ui_dup_method_data.
Add test coverage for CRYPTO_dup_ex_data, use OPENSSL_assert.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2988)
The internals tests for chacha, poly1305 and siphash were erroneously
made conditional on if mdc2 was enabled. Corrected to depend on the
correct algorithms being enabled instead.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2991)
Add a test recipe (test/recipes/15-test_ecparams.t) which uses 'openssl
ecparam' to check the test vectors.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2544)
This involves:
- A directory of valid and invalid PEM-encoded curves.
This is non-exhaustive and can be added to.
- A minor patch to 'openssl ecparam' to make it exit non-zero
when curve validation fails.
- A test recipe is added in a separate commit.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2544)
These were still generated by openssl, but with
the previous commit are corroborated by rustls.
(cherry picked from commit eae1982619)
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2989)
The documentation of this function states that the password parameter
can be NULL. However, the implementation returns an error in this case
due to the inner workings of the HMAC_Init_ex() function.
With this change, NULL password will be treated as an empty string and
PKCS5_PBKDF2_HMAC() no longer fails on this input.
I have also added two new test cases that tests the handling of the
special values NULL and -1 of the password and passlen parameters,
respectively.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1692)
At one point the stack was passing a pointer of the element *before* an
array which is undefined.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2971)
Add ExpectedClientCANames: for client auth this checks to see if the
list of certificate authorities supplied by the server matches the
expected value.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2969)
Check that we handle changes of ciphersuite between HRR and ServerHello
correctly.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2895)
Test that if the server selects a ciphersuite with a different hash from
the PSK in the original ClientHello, the second ClientHello does not
contain the PSK.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2895)
The end of early data is now indicated by a new handshake message rather
than an alert.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2895)
These are self-generated test vectors which gives us very little
confidence that we've got the implementation right. However until
we can get vectors from somewhere else (or ideally official vectors)
this is all we've got. At least it will tell us if we accidentally
break something at some point in the future.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2895)
Add python cryptography testing instructions too
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2885)
Import test data from:
ftp://ftp.rsasecurity.com/pub/pkcs/pkcs-1/pkcs-1v2-1-vec.zip
This is a set of RSA-PSS and RSA-OAEP test vectors including some edge cases
with unusual key sizes.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2944)
The previous 2 commits fixed some issues in the Boring tests. This
re-enables those tests.
[extended tests]
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2942)
After a resumption it is documented that SSL_get_peer_cert_chain() will
return NULL. In BoringSSL it still returns the chain. We don't support that
so we should update the shim to call SSL_get_peer_certificate() instead
when checking whether a peer certificate is available.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2942)
OpenSSL requires that we set the session id context. BoringSSL apparently
does not require this, so wasn't setting it.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2942)
Using a cert with Cyrillic characters, kindly supplied by Dmitry Belyavsky
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2943)
Found using various (old-ish) versions of gcc.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2940)
On some platforms, setting stdout to binary mode isn't quite enough,
which makes the result unusable. With -out, we have better control.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2939)
The Boring runner attempts to enable the NULL-SHA ciphersuite using the
cipherstring "DEFAULT:NULL-SHA". However in OpenSSL DEFAULT permanently
switches off NULL ciphersuites, so we fix this up to be "ALL:NULL-SHA"
instead. We can't change the runner so we have to change the shim to
detect this.
(Merged from https://github.com/openssl/openssl/pull/2933)
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Emilia Käsper <emilia@openssl.org>
We already test DTLS protocol versions. For good measure, add some
DTLS tests with client auth to the new test framework, so that we can
remove the old tests without losing coverage.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Update the list of suppressions so that we can run a later BoringSSL set
of tests. This also adds an ErrorMap to greatly reduce the number of
failing tests. The remaining tests that still fail are just disabled for
now.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2930)
The boring tests are currently failing because they send a PSK extension
which isn't in the last place. This is not allowed in the latest TLS1.3
specs. However the Boring tests we have are based on an old commit that
pre-date when that rule first appeared.
The proper solution is to update the tests to a later commit. But for now
to get travis to go green we disable the failing tests.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2928)
This will make the individual external tests more easily selectable /
deselectable through the usual test selection mechanism.
This also moves external tests to group 95.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2902)
This allows a finer granularity when selecting which tests to run, and
makes the tests more vidible.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2901)
Examples of possible expressions (adapt to your platform):
make test TESTS=-99
make test TESTS=10
make test TESTS=-9?
make test TESTS=-[89]0
make test TESTS=[89]0
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2887)
process. This means no AEAD ciphers and no XTS mode.
Update the test script that uses this output to test cipher suites to not
filter out the now missing cipher modes.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2876)
The value of SSL3_RT_MAX_ENCRYPTED_LENGTH normally includes the compression
overhead (even if no compression is negotiated for a connection). Except in
a build where no-comp is used the value of SSL3_RT_MAX_ENCRYPTED_LENGTH does
not include the compression overhead.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2872)
We just check that if we insert a cookie into an HRR it gets echoed back
in the subsequent ClientHello.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2839)
Change the early data API so that the server must use
SSL_write_early_data() to write to an unauthenticated client.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2737)
This is for consistency with the rest of the API where all the functions
are called *early_data*.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2737)
This is for consistency with the rest of the API where all the functions
are called *early_data*.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2737)
Don't create a custom boolean type for parsing CompressionExpected. Use
the existing one instead.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2814)
- FLAT_INC
- PKCS1_CHECK (the SSL_OP_PKCS1_CHECK options have been
no-oped)
- PKCS_TESTVECT (debugging leftovers)
- SSL_AD_MISSING_SRP_USERNAME (unfinished feature)
- DTLS_AD_MISSING_HANDSHAKE_MESSAGE (unfinished feature)
- USE_OBJ_MAC (note this removes a define from the public header but
very unlikely someone would be depending on it)
- SSL_FORBID_ENULL
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Stephen Henson <steve@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
AGL has a history of pointing out the idiosynchronies/laxness of the
openssl PEM parser in amusing ways. If we want this functionality to
stay present, we should test that it works.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2756)
Generate a fresh certificate and DSA private key in their respective PEM
files. Modify the resulting ASCII in various ways so as to produce input
files that might be generated by non-openssl programs (openssl always
generates "standard" PEM files, with base64 data in 64-character lines
except for a possible shorter last line).
Exercise various combinations of line lengths, leading/trailing
whitespace, non-base64 characters, comments, and padding, for both
unencrypted and encrypted files. (We do not have any other test coverage
that uses encrypted files, as far as I can see, and the parser enforces
different rules for the body of encrypted files.)
Add a recipe to parse these test files and verify that they contain the
expected string or are rejected, according to the expected status.
Some of the current behavior is perhaps suboptimal and could be revisited.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2756)
This removes the fips configure option. This option is broken as the
required FIPS code is not available.
FIPS_mode() and FIPS_mode_set() are retained for compatibility, but
FIPS_mode() always returns 0, and FIPS_mode_set() can only be used to
turn FIPS mode off.
Reviewed-by: Stephen Henson <steve@openssl.org>
The second loop in the remove_space function doesn't check for walking
back off of the start of the string while setting white space to 0.
This fix exits this loop once the pointer is before the (updated) beginning
of the string.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2752)
Avoid a -Wundef warning in refcount.h
Avoid a -Wundef warning in o_str.c
Avoid a -Wundef warning in testutil.h
Include internal/cryptlib.h before openssl/stack.h
to avoid use of undefined symbol OPENSSL_API_COMPAT.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2712)
Make sure that we can stop handshake processing and resume it later.
Also check that the cipher list and compression methods are sane.
Unfortunately, we don't have the client-side APIs needed to force
a specific (known) session ID to be sent in the ClientHello, so
that accessor cannot be tested here.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2279)
Certain callback APIs allow the callback to request async processing
by trickling a particular error value up the stack to the application
as an error return from the handshake function. In those cases,
SSL_want() returns a code specific to the type of async processing
needed.
The create_ssl_connection() helper function for the tests is very
helpful for several things, including creating API tests. However,
it does not currently let us test the async processing functionality
of these callback interfaces, because the special SSL error codes
are treated as generic errors and the helper continues to loop until
it reaches its maximum iteration count.
Add a new parameter, 'want', that indicates an expected/desired
special SSL error code, so that the helper will terminate when
either side reports that error, giving control back to the calling
function and allowing the test to proceed.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2279)
create_ssl_connection() prints out the results if SSL_accept() and/or
SSL_connect() fail, but was reusing the client return value when printing
about SSL_accept() failures.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2279)
Plumb things through in the same place as the SNI callback, since
we recommend that the early callback replace (and supplement) the
SNI callback, and add a few test cases.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2279)
TODO(robpercival): Should actually test that the output certificate
contains the poison extension.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/843)
This implementation is written in endian agnostic C code. No attempt
at providing machine specific assembly code has been made. This
implementation expands the evptests by including the test cases from
RFC 5794 and ARIA official site rather than providing an individual
test case. Support for ARIA has been integrated into the command line
applications, but not TLS. Implemented modes are CBC, CFB1, CFB8,
CFB128, CTR, ECB and OFB128.
Reviewed-by: Andy Polyakov <appro@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2337)
On VMS, file names with more than one period get all but the last get
escaped with a ^, so 21-key-update.conf.in becomes 21-key-update^.conf.in
That means that %conf_dependent_tests and %skip become useless unless
we massage the file names that are used as indexes.
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2678)
Since 20-cert-select.conf will vary depending in no-dh and no-dsa,
don't check it against original when those options are selected
Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2680)
We use an int instead. That means SSL_key_update() also should use an int.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2609)
In 1.1.0 changing the ciphersuite during a renegotiation can result in
a crash leading to a DoS attack. In master this does not occur with TLS
(instead you get an internal error, which is still wrong but not a security
issue) - but the problem still exists in the DTLS code.
The problem is caused by changing the flag indicating whether to use ETM
or not immediately on negotiation of ETM, rather than at CCS. Therefore,
during a renegotiation, if the ETM state is changing (usually due to a
change of ciphersuite), then an error/crash will occur.
Due to the fact that there are separate CCS messages for read and write
we actually now need two flags to determine whether to use ETM or not.
CVE-2017-3733
Reviewed-by: Richard Levitte <levitte@openssl.org>
In 1.1.0 changing the ciphersuite during a renegotiation can result in
a crash leading to a DoS attack. In master this does not occur with TLS
(instead you get an internal error, which is still wrong but not a security
issue) - but the problem still exists in the DTLS code.
This commit provides a test for the issue.
CVE-2017-3733
Reviewed-by: Richard Levitte <levitte@openssl.org>
Make sure we get an HRR in the right circumstances based on kex mode.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)
Previously counting the number of tests in checkhandshake.pm took an
initial guess and then modified it based on various known special
cases. That is becoming increasingly untenable, so this changes it to
properly calculate the number of tests we expect to run.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2341)
This happens when a fd is added and then immediately removed from the
ASYNC_WAIT_CTX before pausing the job.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2581)
test/recipes/40-test_rehash.t uses test files from certs/demo, which
doesn't exist any longer. Have it use PEM files from test/ instead.
Because rehash wants only one certificate or CRL per file, we must
also filter those PEM files to produce test files with a single object
each.
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2594)
If s->s3->tmp.new_cipher is NULL then a crash can occur. This can happen
if an alert gets sent after version negotiation (i.e. we have selected
TLSv1.3 and ended up in tls13_enc), but before a ciphersuite has been
selected.
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2575)
Idea is to keep it last for all eternity, so that if you find yourself
in time-pressed situation and deem that fuzz test can be temporarily
skipped, you can terminate the test suite with less hesitation about
following tests that you would have originally missed.
Reviewed-by: Rich Salz <rsalz@openssl.org>
For TLS 1.3 we select certificates with signature algorithms extension
only. For ECDSA+SHA384 there is the additional restriction that the
curve must be P-384: since the test uses P-256 this should fail.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2339)
