The negative tests accepted any non-zero client status, and the openssl.test
capability and protocol probes read only a command's output. A run killed by
timeout(1) therefore looked like the certificate rejection, auth failure or
missing feature each site was testing for, and the script still reported
success with the coverage silently dropped.
Check for the timeout statuses (124, and 137 for the SIGKILL used here)
before interpreting a result:
- trusted_peer.test: the three wrong-CA / wrong-peer rejection cases
- tls13.test: cipher mismatch, mutual auth, and the version downgrade cases
- psk.test: the no-peer-cert rejection case
- openssl.test: the seven wolfSSL/OpenSSL certificate capability probes and
the SSLv3, TLS 1.0 and TLS 1.1 connection probes
Three ways the bounds added here failed to do their job:
- get_first_free_port ended the scan cap with 'exit 1', but every caller
runs it in a command substitution, so only the subshell died. The port
variable came back empty, the next $((port + 1)) evaluated to 1, and the
run limped on to a confusing wait_for_readyFile failure. Forcing the cap
on ocsp-stapling.test: before, the script ran on and hung until an outer
timeout killed it; now it exits 1 at the error. Return instead, and check
the status at all 25 call sites across the six scripts.
- The macOS timeout shim was a shell function. Backgrounding a function
forks a subshell, so $! was the subshell and cleanup killed that while
the server it was meant to stop leaked. Use a prefix variable that
expands to nothing when timeout(1) is absent, keeping $! the real pid.
- timeout -s KILL exits 137, not 124. Sites that read $? and treat any
non-zero as 'feature not compiled in' turned a hang into exit 0, so the
bound made a hang less visible than before. Add timed_out() and check it
before those skip branches; use it for the version probes too, which
matched any status >= 124.
The 'Bad SSL version' probes branch on grep's status, so a client
killed by the 2m timeout read as 'TLS v1.2 supported' and ran the
wrong branch. Take both statuses from PIPESTATUS and fail loudly on
a probe timeout.
- tls13.test: take the client's exit status from PIPESTATUS[0]; $? after
the pipe was tee's status, hiding a client failure or 2m-timeout kill.
- ocsp-responder-openssl-interop.test: only escalate to SIGKILL if the
responder is still running, and make the reap counter local.
A test script that blocks forever burns the CI job's full
timeout-minutes with no logs. e82ecdff93 and 5c5cbd3094 bounded the
waited-on servers; this covers the remaining hang classes in
scripts/*.test:
- Wrap foreground example client/server, openssl s_client, and
openssl ocsp invocations in "timeout -s KILL 2m". A client wedged
before or without a live peer (e.g. blocked in first-seed entropy
gathering, or DTLS with no reset from a dead peer) is not bounded
by its peer's timeout.
- Add the macOS timeout() fallback shim to scripts that now use
timeout.
- Bound the get_first_free_port scan loops (nc -w 1, 100-port cap).
- Add -w 1 to the remaining nc probes and dtls.test UDP pcap markers.
- ocsp-responder-openssl-interop.test: bound the responder reap in
cleanup: give each responder 5 s to exit after SIGTERM, then
SIGKILL before waiting, so a wedged responder cannot hang the EXIT
trap.
- benchmark.test: bound the clients but leave the -i servers
unwrapped: the script ends them with kill -6, which timeout(1)
does not forward, so wrapping would orphan the server.
- trusted_peer/tls13: kill the server with SIGTERM instead of
SIGKILL in cleanup so the signal forwards through the timeout
wrapper to the wrapped server.
The scripts that wait for a server to publish its ready file declare
counter at file scope and never reset it, so the retry budget is shared
by every server start in the script instead of applying to each one.
Once the early cases have used it up, every later create_port() falls
straight through to "NO ready file ending test", kills a server that was
starting normally, and the client then fails with "port number cannot be
0". Retry loops do not help, since the budget is already spent when they
run.
The failure needs only a build whose server start-up is slow enough to
consume a few tenths of a second each time. It showed up in the FIPS
dev-no-POST kernel-settings-all-pqc-asm job, where the server pays for
the CASTs, the PQC algorithms and the vector-register fallback fuzzer:
psk.test gave up after exactly 20 waits and tls13.test after exactly 51,
both the full script budget rather than a per-case one.
Reset counter where the wait begins, which is what the ocsp-stapling
scripts already do. Reproduced with a wrapper that delays the server by
one second: psk.test then fails on its third case before the change and
passes after it.
timeout(1) is GNU coreutils and is not installed on macOS, so the
"make check macos" job failed with "timeout: command not found" for
every wrapped server. Add a small shim to each affected test: when
timeout is unavailable (e.g. macOS) run the server unbounded, restoring
the prior macOS behavior. The flaky hang the timeout guards against is on
the Linux-only trackmemory job, so macOS does not need the bound.
Several test scripts share the same pattern as ocsp-stapling_tls13multi:
a backgrounded example server is "wait"ed on with no timeout, so a
server that flakily fails to exit blocks the script until the CI job
timeout. Wrap those servers in "timeout -s KILL 2m" as well.
Scripts: ocsp-stapling, ocsp-stapling2,
ocsp-stapling-with-wolfssl-responder, crl-revoked, tls13, resume,
pkcallbacks, dtlscid.
in scripts/tls13.test, use fips-compatible server-side cipher suite for "TLS v1.3 cipher suite mismatch" test, and modernize some syntax;
in configure.ac, omit earlydata from enable-all when asynccrypt, pending fix;
also in configure.ac, fix AC_CHECK_DECLS()-overriding-AC_CHECK_FUNCS() kludge, to fix CPPFLAGS=-std=c99 builds.
Support Draft 28: able to compile code to return BAD_BINDER if no PSKs
match and certificates not to be used.
Change key share implementation to use server preference - server now
checks each client key share's group is in supported_groups extension.
Client and server examples modified to support server preference.
Application can set client's and server's supported groups by rank.
Server's supported groups is sent back in encrypted_extensions if
preferred group is not in client's list - able to be turned off at
compile time.
Application can query server's preferred group from client.
Able to compile using 0x0304 as version instead of draft version.
Fix state machine in TLS 1.3 to support unexpected hello_retry_request.
Also fixes non-blocking.
Fix resumption to use the named group from session.
Fix named group in session structure to be a 2-byte field.
Better detection of errors in message flow.
Fix DoTls13ClientHello when downgrading - don't do TLS 1.3 things.
Not downgrading on client fixed.
Downgrade protocol version from TLS 1.3 when not TLS 1.3 ciphersuite.
Get downgrading from TLS 1.3 and resumption working.
Change earlyData value to an enum.
Support no extensions data (as opposed to zero length extension data) in
TLS 1.3 ClientHello.
Check PSK cipher suite is available to both client and server before
using.
Check first PSK identity chosen when server says it is using early data
at client.
Check PSK extension is last in client_hello on server.
Check the PSK cipher suite to use is supported on client.
Check the returned cipher suite for pre-shared keys is the same as
client expects.
Send alert decrypt_error when verification fails in certificate_verify
or finished message doesn't match calculated value.
Fail when certificate messages recieved in handshake when using PSK.
Validate on the server that EndOfEarlyData message has been recieved
before finished message when server sent EarlyData extension.
Changes made to test.h to allow interop of PSK with OpenSSL.
Changes to allow server to pre-generate key share and perform other
operations at later time.
Fix ChaCha20 code header to have bigger state to support assembly code
for AVX1.
Fix Curve25519 code to use define instead.
Change Curve25519 to memset all object data on init.
Change Poly1305 to put both sizes into one buffer to avoid a second call
to wc_Poly1305Update().
Added WOLFSSL_START and WOLFSSL_END API and calls to show time of
protocol message function enter and leave to analyse performance
differences.
Moved Curve25519 code in KeyShare extension out of general ECC code.
Fix handling of ServerHello in TLS v1.3 to support TLS v1.2 when
downgrading.
Added support in client and server examples for using downgrade method:
wolfSSLv23_client_method_ex() or wolfSSLv23_server_method_ex().
Add tests, using downgrade version, of client or server downgrading from
TLS v1.3 to TLS v1.2.
* Fixes to support certificate generation (`WOLFSSL_CERT_GEN`) without RSA enabled.
* Added new ECC CA for 384-bit tests.
* Created new server cert chain (ECC CA for 256-bit that signs server-ecc.pem)
* Created new `./certs/ecc/genecc.sh` script for generating all ECC CA's, generated server cert req (CSR), signing with CA and the required CRL.
* Moved the wolfCrypt ECC CA / ECC cert gen test into `ecc_test` as `ecc_test_cert_gen`.
* Refactor duplicate code that saves DER to disk, converts DER to PEM and saves PEM to disk into SaveDerAndPem function.
* Changed `ecc_test_make_pub` and `ecc_test_key_gen` to use XMALLOC for temp buffers (uses heap instead of stack).
* Cleanup to combine all certificate subject information into global `certDefaultName`.
* Updated cert request info to use wolfSSL instead of Yassl.
* Cleanup to combine keyUsage into `certKeyUsage` and `certKeyUsage2`.
* Re-number error codes in rsa_test.
* Moved the certext_test after the ecc_test, since it uses a file generated in `ecc_test_cert_gen`.
Change SupportedVersions extension to only include TLS v1.3 if downgrade
is disabled.
Fix parsing of SupportedVersions extension
Don't upgrade
Only downgrade in SupportedVersions extension if option enabled
Changed the define in configure.ac to match the one used in the code.
Fixed downgrading to disallow unless ssl->options.downgrade is set.
TLS 1.3 client method does not have downgrade on anymore.
Test changed to not expect downgrading to work.
Test of TLS v1.3 client downgrade is actually upgrading on server.
Fixed 80 character line problems.
