Conflict resolution:
* ChangeLog
* tests/data_files/Makefile: concurrent additions, order irrelevant
* tests/data_files/test-ca.opensslconf: concurrent additions, order irrelevant
* tests/scripts/all.sh: one comment change conflicted with a code
addition. In addition some of the additions in the
iotssl-1381-x509-verify-refactor-restricted branch need support for
keep-going mode, this will be added in a subsequent commit.
Only delete things that we expect to find, to avoid deleting other
things that people might have lying around in their build tree.
Explicitly skip .git to avoid e.g. accidentally matching a branch
name.
All options can now be overridden by a subsequent option, e.g.
"all.sh --foo --no-foo" is equivalent to "all.sh --no-foo". This
allows making wrapper scripts with default options and occasionally
overriding those options when running the wrapper script.
The 'critical' boolean can be set to false in two ways:
- by leaving it implicit (test data generated by openssl)
- by explicitly setting it to false (generated by hand)
This covers all lines added in the previous commit. Coverage was tested using:
make CFLAGS='--coverage -g3 -O0'
(cd tests && ./test_suite_x509parse)
make lcov
firefox Coverage/index.html # then visual check
Test data was generated by taking a copy of tests/data_files/crl-idp.pem,
encoding it as hex, and then manually changing the values of some bytes to
achieve the desired errors, using https://lapo.it/asn1js/ for help in locating
the desired bytes.
Found by running:
CC=clang cmake -D CMAKE_BUILD_TYPE="Check"
tests/scripts/depend-pkalgs.pl
(Also tested with same command but CC=gcc)
Another PR will address improving all.sh and/or the depend-xxx.pl scripts
themselves to catch this kind of thing.
Our current behaviour is a bit inconsistent here:
- when the bad signature is made by a trusted CA, we stop here and don't
include the trusted CA in the chain (don't call vrfy on it)
- otherwise, we just add NOT_TRUSTED to the flags but keep building the chain
and call vrfy on the upper certs
This ensures that the callback can actually clear that flag, and that it is
seen by the callback at the right level. This flag is not set at the same
place than others, and this difference will get bigger in the upcoming
refactor, so let's ensure we don't break anything here.
When a trusted CA is rolling its root keys, it could happen that for some
users the list of trusted roots contains two versions of the same CA with the
same name but different keys. Currently this is supported but wasn't tested.
Note: the intermediate file test-ca-alt.csr is commited on purpose, as not
commiting intermediate files causes make to regenerate files that we don't
want it to touch.
As we accept EE certs that are explicitly trusted (in the list of trusted
roots) and usually look for parent by subject, and in the future we might want
to avoid checking the self-signature on trusted certs, there could a risk that we
incorrectly accept a cert that looks like a trusted root except it doesn't
have the same key. This test ensures this will never happen.
The tests cover chains of length 0, 1 and 2, with one error, located at any of
the available levels in the chain. This exercises all three call sites of
f_vrfy (two in verify_top, one in verify_child). Chains of greater length
would not cover any new code path or behaviour that I can see.
So far there was no test ensuring that the flags passed to the vrfy callback
are correct (ie the flags for the current certificate, not including those of
the parent).
Actual tests case making use of that test function will be added in the next
commit.
With cmake, CFLAGS has to be set when invoking cmake, not make (which totally
ignores the value of CFLAGS when it runs and only keeps the one from cmake).
Also, in that case the flags were either redundant (-Werror etc) or wrong
(-std=c99 -pedantic) as some parts of the library will not build with
-pedantic (see the other -pedantic tests, which are correct, for what needs to
be disabled).
This is step 1 of a plan to get rid once and for all of missing depends_on in
the X509 test suite (step 2 will be RSA/ECDSA, and step 0 was curves.pl).
We have code to skip them but didn't have explicit tests ensuring they are
(the corresponding branch was never taken).
While at it, remove extra copy of the chain in server10*.crt, which was
duplicated for no reason.
This shows inconsistencies in how flags are handled when callback fails:
- sometimes the flags set by the callback are transmitted, sometimes not
- when the cert if not trusted, sometimes BADCERT_NOT_TRUSTED is set,
sometimes not
This adds coverage for 9 lines and 9 branches. Now all lines related to
callback failure are covered.
Now all checks related to profile are covered in:
- verify_with_profile()
- verify_child()
- verify_top()
(that's 10 lines that were previously not covered)
Leaving aside profile enforcement in CRLs for now, as the focus is on
preparing to refactor cert verification.
Previously flags was left to whatever value it had before. It's cleaner to
make sure it has a definite value, and all bits set looks like the safest way
for when it went very wrong.
Build with MBEDTLS_DEPRECATED_REMOVED and MBEDTLS_DEPRECATED_WARNING
separately.
Do these builds with `-O -Werror -Wall -Wextra` to catch a maximum of
issues while we're at it. Do one with gcc and one with clang for
variety. This caught an uninitialized variable warning in cmac.c that
builds without -O didn't catch.
Extend the pkparse test suite with the newly created keys
encrypted using PKCS#8 with PKCS#5 v2.0 with PRF being
SHA224, 256, 384 and 512.
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
We now have support for the entire SHA family to be used as
PRF in PKCS#5 v2.0, therefore we need to add new keys to test
these new functionalities.
This patch adds the new keys in `tests/data_files` and
commands to generate them in `tests/data_files/Makefile`.
Note that the pkcs8 command in OpenSSL 1.0 called with
the -v2 argument generates keys using PKCS#5 v2.0 with SHA1
as PRF by default.
(This behaviour has changed in OpenSSL 1.1, where the exact same
command instead uses PKCS#5 v2.0 with SHA256)
The new keys are generated by specifying different PRFs with
-v2prf.
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Some unit tests for pbkdf2_hmac() have results longer than
99bytes when represented in hexadecimal form.
For this reason extend the result array to accommodate
longer strings.
At the same time make memset() parametric to avoid
bugs in the future.
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Test vectors for SHA224,256,384 and 512 have been
generated using Python's hashlib module by the
following oneliner:
import binascii, hashlib
binascii.hexlify(hashlib.pbkdf2_hmac(ALGO, binascii.unhexlify('PASSWORD'), binascii.unhexlify('SALT'), ITER, KEYLEN)))
where ALGO was 'sha224', 'sha256', 'sha384' and 'sha512'
respectively.
Values for PASSWORD, SALT, ITER and KEYLEN were copied from the
existent test vectors for SHA1.
For SHA256 we also have two test vectors coming from RFC7914 Sec 11.
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Currently only SHA1 is supported as PRF algorithm for PBKDF2
(PKCS#5 v2.0).
This means that keys encrypted and authenticated using
another algorithm of the SHA family cannot be decrypted.
This deficiency has become particularly incumbent now that
PKIs created with OpenSSL1.1 are encrypting keys using
hmacSHA256 by default (OpenSSL1.0 used PKCS#5 v1.0 by default
and even if v2 was forced, it would still use hmacSHA1).
Enable support for all the digest algorithms of the SHA
family for PKCS#5 v2.0.
Signed-off-by: Antonio Quartulli <antonio@openvpn.net>
Build with MBEDTLS_HAVE_INT32 and MBEDTLS_HAVE_INT64 on all
architectures, not just x86_64. These two modes should work on all
platforms (except embedded environments where 64-bit division is not
available).
Also run the unit tests.
Correct the description: this is not "N-bit compilation", but "N-bit
bignum limbs".
This change fixes a problem in the tests pk_rsa_alt() and
pk_rsa_overflow() from test_suite_pk.function that would cause a
segmentation fault. The problem is that these tests are only designed
to run in computers where the SIZE_MAX > UINT_MAX.
Change function in tests named mbedtls_entropy_func to
mbedtls_test_entropy_func to avoid getting error from the linker when
calling the mbedtls_entropy_func elsewhere.
The _ext suffix suggests "new arguments", but the new functions have
the same arguments. Use _ret instead, to convey that the difference is
that the new functions return a value.
If lsof is not available, wait_server_start uses a fixed timeout,
which can trigger a race condition if the timeout turns out to be too
short. Emit a warning so that we know this is going on from the test
logs.
- Some of the CI machines don't have lsof installed yet, so rely on an sleeping
an arbitrary number of seconds while the server starts. We're seeing
occasional failures with the current delay because the CI machines are highly
loaded, which seems to indicate the current delay is not quite enough, but
hopefully not to far either, so double it.
- While at it, also double the watchdog delay: while I don't remember seeing
much failures due to client timeout, this change doesn't impact normal
running time of the script, so better err on the safe side.
These changes don't affect the test and should only affect the false positive
rate coming from the test framework in those scripts.
Conflict resolution:
* ChangeLog: put the new entries in their rightful place.
* library/x509write_crt.c: the change in development was whitespace
only, so use the one from the iotssl-1251 feature branch.
1) `mbedtls_rsa_import_raw` used an uninitialized return
value when it was called without any input parameters.
While not sensible, this is allowed and should be a
succeeding no-op.
2) The MPI test for prime generation missed a return value
check for a call to `mbedtls_mpi_shift_r`. This is neither
critical nor new but should be fixed.
3) Both the RSA keygeneration example program and the
RSA test suites contained code initializing an RSA context
after a potentially failing call to CTR DRBG initialization,
leaving the corresponding RSA context free call in the
cleanup section of the respective function orphaned.
While this defect existed before, Coverity picked up on
it again because of newly introduced MPI's that were
also wrongly initialized only after the call to CTR DRBG
init. The commit fixes both the old and the new issue
by moving the initializtion of both the RSA context and
all MPI's prior to the first potentially failing call.
- Adapt the change in all.sh to the new keep-going mode
- Restore alphabetical order of configuration flags for
alternative implementations in config.h and rebuild
library/version_features.c
Compilation and test for the `MBEDTLS_RSA_NO_CRT` option were
previously guarded by a check for 64-bit systems, for which there
is no reason. This commit moves both outside of the guard.
* development:
all.sh: add some documentation
all.sh: new option --no-armcc
all.sh: add --yotta to go with --no-yotta
all.sh: --keep-going mode
all.sh: cleaned up usage output
all.sh: indent
* public/pr/1207:
all.sh: add some documentation
all.sh: new option --no-armcc
all.sh: add --yotta to go with --no-yotta
all.sh: --keep-going mode
all.sh: cleaned up usage output
all.sh: indent
* development:
Timing self test: shorten redundant tests
Timing self test: increased duration
Timing self test: increased tolerance
Timing unit tests: more protection against infinite loops
Unit test for mbedtls_timing_hardclock
New timing unit tests
selftest: allow excluding a subset of the tests
selftest: allow running a subset of the tests
selftest: refactor to separate the list of tests from the logic
Timing self test: print some diagnosis information
mbedtls_timing_get_timer: don't use uninitialized memory
timing interface documentation: minor clarifications
Timing: fix mbedtls_set_alarm(0) on Unix/POSIX
* public/pr/1136:
Timing self test: shorten redundant tests
Timing self test: increased duration
Timing self test: increased tolerance
Timing unit tests: more protection against infinite loops
Unit test for mbedtls_timing_hardclock
New timing unit tests
selftest: allow excluding a subset of the tests
selftest: allow running a subset of the tests
selftest: refactor to separate the list of tests from the logic
Timing self test: print some diagnosis information
mbedtls_timing_get_timer: don't use uninitialized memory
timing interface documentation: minor clarifications
Timing: fix mbedtls_set_alarm(0) on Unix/POSIX
* Correct order of sections in ChangeLog
* Restore unintentionally removed whitespace and
formatting improvements.
* Consistently rename MBEDTLS_ERR_RSA_EXPORT_UNSUPPORTED
to MBEDTLS_ERR_RSA_UNSUPPORTED_OPERATION in rsa.h
documentation.
During the work on the RSA change the issue was brought up,
and a fix was provided on development, that some RSA tests
use CTR DRBG and depend on the presence of a strong entropy
source to succeed. The RSA work introduced more tests using
CTR DRBG, and the dependency needs to be added for them, too.
Add --keep-going mode to all.sh. In this mode, if a test fails, keep
running the subsequent tests. If a build fails, skip any tests of this
build and move on to the next tests. Errors in infrastructure, such as
git or cmake runs, remain fatal. Print an error summary at the end of
the run, and return a nonzero code if there was any failure.
In known terminal types, use color to highlight errors.
On a fatal signal, interrupt the run and report the errors so far.
If timing_timer_simple fails because it detects that timers are likely
to never expire (e.g. going backward or not incrementing), skip all
tests that rely on timers.
Do test mbedtls_timing_hardclock. We can't reliably test much about
it, but at least test that it doesn't crash, isn't constant, and
doesn't look completely random.
New set of unit tests for the timing module, instead of just running
the selftest function.
The selftest function sometimes fails on a heavily loaded
machine (such as a typical continuous integration system). Because of
the all-in-one nature of the test and because the exact load pattern
can be hard to reproduce, it is difficult to diagnose failures of CI
runs with selftest. The new tests are more separated and I strove to
point out potential failure modes in comments.
* mbedtls_timing_hardclock: not tested. This function gives so few
guarantees that there isn't much to test, and it is hard to test
reliably because clock cycles don't easily relate to time in any
remotely portable way. This function isn't used in the library
anyway, it's only there for benchmark programs.
* mbedtls_timing_get_timer: tested by setting a timer and verifying
that it reaches its target, and by verifying that a timer started
later than another always has a smaller elapsed time.
* mbedtls_set_alarm: tested by setting an alarm, busy-waiting for it
and measuring the elapsed time with a timer.
* mbedtls_timing_set_delay, mbedtls_timing_get_delay: tested by
setting a delay object and watching it go through its two delay
values, using a timer to check that the delays are passed at the
expected time.
The tests pass under light to moderate load, but some of them can be
defeated with sufficiently heavy load. This is unavoidable since the
test process to be effectively suspended for any length of time,
making us think that a timer has gone on for too long.
* development:
Don't split error code description across multiple lines
Register new error code in error.h
Move deprecation to separate section in ChangeLog
Extend scope of ERR_RSA_UNSUPPORTED_OPERATION error code
Adapt RSA test suite
Adapt ChangeLog
Deprecate usage of RSA primitives with wrong key type
* restricted/pr/397:
Don't split error code description across multiple lines
Register new error code in error.h
Move deprecation to separate section in ChangeLog
Extend scope of ERR_RSA_UNSUPPORTED_OPERATION error code
Adapt RSA test suite
Adapt ChangeLog
Deprecate usage of RSA primitives with wrong key type
Port wait_server_start from ssl-opt.sh to compat.sh, instead of just
using "sleep 1". This solves the problem that on a heavily loaded
machine, sleep 1 is sometimes not enough (we had CI failures because
of this). This is also faster on a lightly-loaded machine (execution
time reduced from ~8min to ~6min on my machine).
In wait_server_start, fork less. When lsof is present, call it on the
expected process. This saves a few percent of execution time on a
lightly loaded machine. Also, sleep for a short duration rather than
using a tight loop.
