fix(gnutls): replace UDP test skip with readiness polling

Wait for the UDP socket to bind before starting the client, detect
server exit, and clean up on skip. Preserve both handshake checks.

Add regression checks and document the rationale for upstream submission.

Validated on x86-64-v3: 796 passes, 131 existing skips, zero failures.
All seven focused checks and the delayed-start reproduction passed.
This commit is contained in:
2026-09-18 12:08:10 -04:00
parent 3614428e3d
commit 5a59c4d0f5
7 changed files with 429 additions and 17 deletions
+1
View File
@@ -15,6 +15,7 @@
}; };
}; };
patches = import ./patches;
test-exclusions = import ./test-exclusions.nix; test-exclusions = import ./test-exclusions.nix;
x86-64-v3-workarounds = import ./x86-64-v3-workarounds.nix; x86-64-v3-workarounds = import ./x86-64-v3-workarounds.nix;
+3
View File
@@ -0,0 +1,3 @@
_final: prev: {
gnutls = import ./gnutls { inherit (prev) gnutls; };
}
+160
View File
@@ -0,0 +1,160 @@
# GnuTLS UDP server readiness
Under load, the test client can start before `gnutls-serv` binds its UDP
socket, and the first handshake fails with `Connection refused`.
`serv-udp.sh` currently waits a fixed four seconds; elapsed time does not
establish server readiness. `udp-server-readiness.patch` replaces that wait
with polling for the local IPv4 UDP endpoint.
## Scope and waiting behavior
The patch changes the existing `wait_udp_server()` and adds a new
`check_if_udp_port_bound()` beside it in `tests/scripts/common.sh`.
`serv-udp.sh` is its only caller in 3.8.13. The TCP helpers `wait_server()`
and `wait_for_port()`, including their existing sleeps, are unchanged.
Both original DTLS handshake checks remain unchanged.
Each iteration checks process liveness and the socket **before sleeping**.
A ready socket returns immediately. An unsuccessful check sleeps two
seconds only if another attempt remains: at most 90 attempts, consistent
with the existing `wait_server()` budget implemented by `wait_for_port()`,
with no sleep after the final check. Server exit fails early; exhausting the
budget fails and terminates the server. No handshake is retried, and no
protocol timeout is changed. Once bound, the kernel can queue datagrams
while the server is scheduled; the probe itself sends no packets.
The existing `have_port_finder()` prefers `ss`, then `netstat`. If neither
exists, it prints `neither ss nor netstat found` and exits **77 (skip)**.
In the normal test flow, port selection calls it before launching a server.
The probe runs in a subshell so that, even if this skip occurs after launch,
the waiting helper can terminate and reap the server before exiting 77.
## Why an IPv4 socket is expected
This is specific to the server used by this test, not a general rule that
IPv6 sockets cannot serve IPv4 clients. The client explicitly uses
`127.0.0.1`. The server's `--udp` path calls `udp_server()`, which calls
`listen_socket(..., SOCK_DGRAM)`. That function iterates the wildcard
addresses returned by `getaddrinfo(NULL, ..., AI_PASSIVE)`:
| Server build / Linux setting | Binding behavior |
| --- | --- |
| IPv6 enabled, `net.ipv6.bindv6only=0` | Requests `IPV6_V6ONLY=1` on the IPv6 socket, binds `[::]:PORT`, and separately binds `0.0.0.0:PORT`. It overrides the system's dual-stack default. |
| IPv6 enabled, `net.ipv6.bindv6only=1` | The same explicit socket option and separate IPv4/IPv6 binds. |
| `HAVE_IPV6` undefined | Skips every address family except `AF_INET`; only the IPv4 wildcard is attempted. |
`udp_server()` uses `wait_for_connection()`, which puts **every listener**
from that list into `select()` and returns a readable socket for `recvfrom()`;
it does not permanently choose one socket based on `getaddrinfo()` order.
The first two cases were traced with the actual GnuTLS 3.8.13 binary in
separate Linux network namespaces: `setsockopt(IPV6_V6ONLY, [1])` and both
UDP binds returned success under each setting. The no-IPv6 case was checked
in source, not by building a second binary. The same bind implementation
was checked directly on GitLab master.
Thus, successful normal startup for this invocation provides an explicit
IPv4 socket; a lone IPv6 wildcard is not the expected success path.
There is one portability caveat: upstream discards the return value of
`setsockopt(IPV6_V6ONLY)`. On a platform where that call fails and the server
ends up with only a dual-stack socket, this helper would time out despite
IPv4 reachability. Such a platform needs additional handling before this
patch can claim support. Blindly accepting every IPv6 wildcard would also
accept IPv6-only sockets before the separate IPv4 bind finishes.
Source: [`src/serv.c`, `listen_socket()`](https://gitlab.com/gnutls/gnutls/-/blob/master/src/serv.c#L937),
[`src/udp-serv.c`](https://gitlab.com/gnutls/gnutls/-/blob/master/src/udp-serv.c),
and [`tests/serv-udp.sh`](https://gitlab.com/gnutls/gnutls/-/blob/master/tests/serv-udp.sh).
## Port matching and ownership limit
Only `-an` is passed to the socket-listing tool: BSD `netstat -u` selects
Unix-domain sockets, whereas Linux `netstat -u` selects UDP. The parser
handles the extra state column in `ss`, Linux colon-separated endpoints,
and BSD dot-separated endpoints, including `*.PORT`. It matches the full
local port and rejects TCP, IPv6 entries, peer ports, and longer numbers.
A live PID plus a bound port does **not** prove that PID owns the socket.
Existing `GETPORT` selection checks for an unused port and uses a test
port-lock directory; `launch_bare_server()` also calls
`wait_for_free_port()` before starting the process. These are advisory:
the launcher does not enforce the latter's result, and another process
can bind between the check and launch. The patch does not close that race
or add nonportable PID parsing. An unrelated process can satisfy the
socket check; the real handshakes remain the functional check and may
fail (or reach the wrong server). This is a startup-order fix, not a
socket-ownership guarantee.
## Reproduction and focused checks
Apply the patch to an unpacked source tree, then run the companion checks
with Python's standard library and a shell:
```sh
patch --fuzz=0 -d /path/to/gnutls -p1 < udp-server-readiness.patch
SHELL=/bin/sh python3 verify-readiness.py /path/to/gnutls/tests/scripts/common.sh -v
```
Set `NETSTAT=/path/to/netstat` to exercise one outside `PATH`. The checks
cover Linux/BSD output samples, false matches, immediate readiness,
missing tools, process exit, timeout cleanup, and real IPv4 UDP sockets
whose bind is delayed six seconds. The missing-tools fixture is skipped
if an absolute fallback `ss` path cannot be hidden with `PATH`. Native
BSD execution remains untested.
To reproduce with GnuTLS itself, run `tests/serv-udp.sh` with `SERV` pointing
to a wrapper that sleeps six seconds, then `exec`s `gnutls-serv` with all
arguments. Set `CLI` to the matching `gnutls-cli`, `srcdir` to the source
`tests` directory, and `abs_top_builddir` to a writable build directory.
With GnuTLS 3.8.13, the original helper failed the first handshake with
`Connection refused`; the patched helper passed both with the same binaries.
## GnuTLS submission
Development and merge requests are on [GitLab](https://gitlab.com/gnutls/gnutls).
[`CONTRIBUTING.md` on master](https://gitlab.com/gnutls/gnutls/-/blob/master/CONTRIBUTING.md)
was read directly for this review. It requires the contributor's DCO
`Signed-off-by`, successful and failure test coverage, consistent coding
style, and adequate documentation; GitLab CI runs for merge requests.
Its commenting guidance asks for comments explaining non-obvious behavior
or protocol expectations. It does not prescribe an additional special
test-suite comment. The patch now explains its IPv4 binding assumption
next to the probe.
The submission will contain the shell patch, without the Python verifier
or a new Python test dependency. The existing `serv-udp.sh` supplies the
functional success check. Running it through the six-second startup
wrapper supplies a reproducible regression case: it fails before the fix
and passes after it. The local verifier was used to validate socket-output
parsing and the helper's success, process-exit, skip-cleanup, and timeout
branches. Those branch checks are local evidence, not new automated
coverage in the upstream suite; the MR must state that distinction.
No dedicated unit-test harness for these shell helpers was found in the
3.8.13 tests inspected. That does not establish that Python cannot be used
upstream; keeping this submission dependency-free is a scope choice. Use
the existing test and before/after reproduction as the submission's
coverage argument, retaining the platform limitations above. Apply the
patch in an upstream checkout and include those results with the
contributor's own sign-off. No MR or sign-off has been created.
## Local NixOS integration and build results
`overlays/default.nix` imports the `overlays/patches` overlay, which loads
`gnutls/default.nix` to apply the patch and keep `serv-udp.sh` enabled.
The patch itself has no Nix dependencies and applies to 3.8.13 and GitLab
master without fuzz.
The final patch was rebuilt with:
```sh
nix build --no-link -L .#nixosConfigurations.jeeves.pkgs.gnutls
```
That x86-64-v3 build passed: 927 tests, 796 passes, 131 existing skips,
zero failures/errors, and `PASS: serv-udp.sh`. The patch bytes in the built
derivation were compared with the repository artifact; both have SHA-256
`59013d47fd446f2dd065012a2259ccc1898fedc8a053a630e13efa0076368760`.
All seven local checks passed, including skip cleanup and exactly 90
probes with 89 sleeps on timeout. The six-second before/after reproduction
was also repeated successfully with the final helper.
+6
View File
@@ -0,0 +1,6 @@
{ gnutls }:
gnutls.overrideAttrs (old: {
# Keep the UDP handshake test enabled on loaded builders by waiting for
# the server to bind its socket. Kept as a standalone patch for upstream.
patches = (old.patches or [ ]) ++ [ ./udp-server-readiness.patch ];
})
@@ -0,0 +1,70 @@
Subject: [PATCH] tests: wait for the UDP server socket before connecting
A fixed four-second sleep does not guarantee that gnutls-serv has bound
its UDP socket on a busy builder. Poll the local IPv4 UDP endpoint using
the existing ss/netstat discovery, with the same retry budget as the TCP
helper. Fail early if the server exits, and retain the original handshake
checks in serv-udp.sh.
Use flags common to ss and BSD/Linux netstat. Match the local endpoint
and complete port number, excluding TCP, IPv6-only and peer endpoints.
--- a/tests/scripts/common.sh
+++ b/tests/scripts/common.sh
@@ -185,10 +185,55 @@
fi
}
+check_if_udp_port_bound() {
+ local PORT=$1
+ have_port_finder
+ # Use only -an, which is shared by ss and BSD/Linux netstat. UDP has
+ # no LISTEN state. Match the local IPv4 endpoint, not a peer port or
+ # a longer port number. serv-udp.sh connects to 127.0.0.1;
+ # listen_socket() in serv.c binds IPv4 separately and requests
+ # IPV6_V6ONLY=1 for its IPv6 socket.
+ $PFCMD -an | awk -v port="$PORT" '
+ $1 == "udp" || $1 == "udp4" {
+ # ss includes a state column; netstat does not.
+ address = ($2 == "UNCONN" || $2 == "ESTAB") ? $5 : $4
+ if (address ~ ("^[0-9.]+[.:]" port "$") ||
+ address == "*." port)
+ found = 1
+ }
+ END { exit !found }
+ '
+}
+
wait_udp_server() {
local PID=$1
+ local ret
trap "test -n \"${PID}\" && kill ${PID};exit 1" 1 15 2
- sleep 4
+ local i=0
+ # Use the same retry budget as wait_for_port(), but also stop if the
+ # server exits before binding its socket.
+ while test $i -lt 90; do
+ if ! kill -0 "$PID" 2>/dev/null; then
+ fail "" "UDP server $PID exited before binding port $PORT"
+ fi
+ # Contain have_port_finder's exit so a skip also stops the server.
+ if (check_if_udp_port_bound "$PORT"); then
+ return 0
+ else
+ ret=$?
+ if test "$ret" = 77; then
+ kill "$PID" 2>/dev/null || :
+ wait "$PID" 2>/dev/null || :
+ exit 77
+ fi
+ fi
+ i=$((i + 1))
+ if test $i -lt 90; then
+ echo "try $i: waiting for UDP port $PORT"
+ sleep 2
+ fi
+ done
+ fail "$PID" "UDP server $PORT did not come up"
}
create_testdir() {
+180
View File
@@ -0,0 +1,180 @@
#!/usr/bin/env python3
"""Exercise patched common.sh without building GnuTLS (Python standard library only).
Usage: python3 verify-readiness.py /path/to/patched/tests/scripts/common.sh
Set SHELL to test another shell, and NETSTAT to test a netstat outside PATH.
"""
# Use unittest so this upstream companion tool needs no pytest installation.
# ruff: noqa: PT009
import os
import shutil
import socket
import subprocess
import sys
import tempfile
import time
import unittest
from pathlib import Path
COMMON = str(Path(sys.argv.pop(1)).resolve())
SHELL = os.environ.get("SHELL", "/bin/sh")
class ReadinessTests(unittest.TestCase):
"""Check endpoint parsing and the server startup lifecycle."""
def setUp(self) -> None:
"""Create a socket-listing fixture for each check."""
self.tmp = tempfile.TemporaryDirectory()
self.addCleanup(self.tmp.cleanup)
self.root = Path(self.tmp.name)
self.fixture = self.root / "sockets"
self.fixture.write_text("")
self.finder = self.root / "port-finder"
self.finder.write_text('#!/bin/sh\ncat "$SOCKET_FIXTURE"\n')
self.finder.chmod(0o755)
def run_shell(self, body: str, **env: str) -> subprocess.CompletedProcess[str]:
"""Source the actual helper and run a shell scenario."""
return subprocess.run(
[SHELL, "-c", '. "$COMMON"\n' + body],
env={
**os.environ,
"COMMON": COMMON,
"SOCKET_FIXTURE": str(self.fixture),
"PFCMD": str(self.finder),
"PORT": "12345",
**env,
},
capture_output=True,
text=True,
timeout=20,
check=False,
)
def test_socket_formats_and_false_matches(self) -> None:
"""Accept IPv4 UDP local endpoints and reject unrelated sockets."""
cases = [
("udp UNCONN 0 0 0.0.0.0:12345 0.0.0.0:*", True),
("udp UNCONN 0 0 127.0.0.1:12345 0.0.0.0:*", True),
("udp 0 0 0.0.0.0:12345 0.0.0.0:*", True),
("udp4 0 0 *.12345 *.*", True),
("udp 0 0 127.0.0.1.12345 *.*", True),
("udp 0 0 *.12345 *.*", True),
("udp UNCONN 0 0 0.0.0.0:123456 0.0.0.0:*", False),
("udp 0 0 0.0.0.0:123456 0.0.0.0:*", False),
("udp ESTAB 0 0 127.0.0.1:54321 127.0.0.1:12345", False),
("udp 0 0 127.0.0.1:54321 127.0.0.1:12345", False),
("tcp LISTEN 0 128 0.0.0.0:12345 0.0.0.0:*", False),
("tcp 0 0 0.0.0.0:12345 0.0.0.0:* LISTEN", False),
("udp UNCONN 0 0 [::]:12345 [::]:*", False),
("udp UNCONN 0 0 *:12345 *:*", False),
("udp6 0 0 :::12345 :::*", False),
("udp6 0 0 *.12345 *.*", False),
("", False),
]
for row, ready in cases:
with self.subTest(row=row):
self.fixture.write_text(row + "\n")
result = self.run_shell('check_if_udp_port_bound "$PORT"')
self.assertEqual(result.returncode, 0 if ready else 1, result.stderr)
def test_exited_server_fails_immediately(self) -> None:
"""Fail without sleeping when the server has already exited."""
result = self.run_shell(
'true &\npid=$!\nwait "$pid"\nsleep() { echo "unexpected sleep" >&2; }\nwait_udp_server "$pid"'
)
self.assertEqual(result.returncode, 1)
self.assertIn("exited before binding", result.stderr)
self.assertNotIn("unexpected sleep", result.stderr)
def test_ready_socket_does_not_sleep(self) -> None:
"""Check readiness before the first sleep."""
self.fixture.write_text("udp UNCONN 0 0 0.0.0.0:12345 0.0.0.0:*\n")
result = self.run_shell('sleep() { echo "unexpected sleep" >&2; }\nwait_udp_server "$$"')
self.assertEqual(result.returncode, 0, result.stderr)
self.assertNotIn("unexpected sleep", result.stderr)
def test_missing_port_finders_skip(self) -> None:
"""Skip and stop the live server when no finder is available."""
# have_port_finder also tries these paths independently of PATH.
if any(os.access(f"{directory}/ss", os.X_OK) for directory in ("/sbin", "/usr/sbin", "/usr/local/sbin")):
self.skipTest("an absolute ss path cannot be hidden by this PATH-only fixture")
with subprocess.Popen(["sleep", "60"]) as server:
try:
result = self.run_shell(
'unset PFCMD\nPATH=/nonexistent\nwait_udp_server "$SERVER_PID"',
SERVER_PID=str(server.pid),
)
self.assertEqual(result.returncode, 77)
self.assertIn("neither ss nor netstat found", result.stderr)
server.wait(timeout=3)
self.assertLess(server.returncode, 0)
finally:
if server.poll() is None:
server.kill()
def test_timeout_is_bounded_and_cleans_up(self) -> None:
"""Stop polling after the retry budget and terminate the server."""
# Only accelerate the polling delay; keep a real live server process.
self.finder.write_text('#!/bin/sh\necho probe >&2\ncat "$SOCKET_FIXTURE"\n')
with subprocess.Popen(["sleep", "60"]) as server:
try:
result = self.run_shell(
'sleep() { echo polling-sleep; }\nwait_udp_server "$SERVER_PID"',
SERVER_PID=str(server.pid),
)
self.assertEqual(result.returncode, 1)
self.assertIn("did not come up", result.stderr)
self.assertEqual(result.stderr.count("probe\n"), 90)
self.assertEqual(result.stdout.count("polling-sleep"), 89)
server.wait(timeout=3)
self.assertLess(server.returncode, 0)
finally:
if server.poll() is None:
server.kill()
def test_server_exits_while_waiting(self) -> None:
"""Detect a startup failure that happens after polling begins."""
result = self.run_shell('sleep 1 &\npid=$!\nwait_udp_server "$pid"')
self.assertEqual(result.returncode, 1)
self.assertIn("exited before binding", result.stderr)
self.assertIn("waiting for UDP port", result.stdout)
def test_real_socket_delayed_beyond_four_seconds(self) -> None:
"""Wait for a real delayed bind with each installed port finder."""
finders = [shutil.which("ss"), os.environ.get("NETSTAT") or shutil.which("netstat")]
finders = [finder for finder in finders if finder]
if not finders:
self.skipTest("neither ss nor netstat available")
for finder in finders:
with self.subTest(finder=finder):
with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as sock:
sock.bind(("127.0.0.1", 0))
port = sock.getsockname()[1]
code = (
"import socket,time,sys; time.sleep(6); "
"s=socket.socket(socket.AF_INET,socket.SOCK_DGRAM); "
"s.bind(('127.0.0.1',int(sys.argv[1]))); time.sleep(30)"
)
with subprocess.Popen([sys.executable, "-c", code, str(port)]) as server:
try:
started = time.monotonic()
result = self.run_shell(
'wait_udp_server "$SERVER_PID"',
SERVER_PID=str(server.pid),
PORT=str(port),
PFCMD=finder,
)
self.assertEqual(result.returncode, 0, result.stderr)
self.assertGreaterEqual(time.monotonic() - started, 6)
self.assertIsNone(server.poll())
finally:
server.terminate()
server.wait(timeout=3)
if __name__ == "__main__":
unittest.main()
+9 -17
View File
@@ -9,14 +9,6 @@
# architecture-dependent floating-point differences whose risk we accept for # architecture-dependent floating-point differences whose risk we accept for
# our workloads. Keep these exceptions visible until their causes are fixed. # our workloads. Keep these exceptions visible until their causes are fixed.
_final: prev: { _final: prev: {
gnutls = prev.gnutls.overrideAttrs (old: {
# This test uses a fixed four-second sleep instead of checking UDP
# readiness; the client saw no listener in the x86-64-v3 build.
postPatch = (old.postPatch or "") + ''
sed '2iexit 77' -i tests/serv-udp.sh
'';
});
prometheus = prev.prometheus.overrideAttrs ( prometheus = prev.prometheus.overrideAttrs (
old: old:
let let
@@ -83,15 +75,15 @@ _final: prev: {
''; '';
}); });
scipy = pythonPrev.scipy.overridePythonAttrs (old: { # scipy = pythonPrev.scipy.overridePythonAttrs (old: {
# x86-64-v3 FFT implementations produce rounding differences outside # # x86-64-v3 FFT implementations produce rounding differences outside
# these tests' strict tolerances. We accept the numerical-precision # # these tests' strict tolerances. We accept the numerical-precision
# risk for our workloads. # # risk for our workloads.
disabledTests = (old.disabledTests or [ ]) ++ [ # disabledTests = (old.disabledTests or [ ]) ++ [
"test_roundtrip_float32" # "test_roundtrip_float32"
"test_roundtrip_scaling" # "test_roundtrip_scaling"
]; # ];
}); # });
sentry-sdk = pythonPrev.sentry-sdk.overridePythonAttrs (old: { sentry-sdk = pythonPrev.sentry-sdk.overridePythonAttrs (old: {
# This test globally mocks threading.current_thread while another # This test globally mocks threading.current_thread while another