extmod/asyncio/stream.py: Add ipv6 support to start_server(). #17311

andrewleech · 2025-05-16T01:19:38Z

Summary

I'm starting to use ipv6 more with micropython devices as their discoverability can actually be much better on a local network, without worrying about needing to manually configure (ipv4) ip addresses.

To that end, this PR adds support to start_server() to automatically use the parsed protocol, ie setting socket.AF_INET or socket.AF_INET6 on the socket as parsed by getaddrinfo()

When adding unit tests for this change, I found the multi-test runner was missing support for ipv6 ip address detection, so that's been added as well (in a separate commit)

The ip address detection in multi-test runner has also been updated to use the newer / preferred network.ipconfig("addr4") style interface by default, falling back to ipconfig if needed.

Testing

Unit tests are included which check ipv6 support in both net_hosted style test and mult_net tests.

Without the change to extmod/asyncio/stream.py the tests fails with OSError: 97

Trade-offs and Alternatives

codecov · 2025-05-16T01:21:56Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.38%. Comparing base (2dada06) to head (6684282).
⚠️ Report is 311 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #17311      +/-   ##
==========================================
- Coverage   98.54%   98.38%   -0.16%     
==========================================
  Files         169      171       +2     
  Lines       21898    22257     +359     
==========================================
+ Hits        21579    21898     +319     
- Misses        319      359      +40

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

github-actions · 2025-05-16T01:35:20Z

Code size report:

   bare-arm:    +0 +0.000% 
minimal x86:    +0 +0.000% 
   unix x64:   +56 +0.007% standard
      stm32:   +24 +0.006% PYBV10
     mimxrt:   +24 +0.006% TEENSY40
        rp2:   +16 +0.002% RPI_PICO_W
       samd:   +24 +0.009% ADAFRUIT_ITSYBITSY_M4_EXPRESS
  qemu rv32:    +0 +0.000% VIRT_RV32

dpgeorge · 2025-07-09T06:16:13Z

extmod/asyncio/stream.py

@@ -180,11 +180,11 @@ async def start_server(cb, host, port, backlog=5, ssl=None):
    import socket

    # Create and bind server socket.
-    host = socket.getaddrinfo(host, port)[0]  # TODO this is blocking!
-    s = socket.socket()
+    addr_info = socket.getaddrinfo(host, port)[0]  # TODO this is blocking!


It would be possible to add the family argument to start_server() and pass it through to getaddrinfo() here. Then you could force it to use IPv6 (or IPv4). But maybe that's a separate PR?

dpgeorge · 2025-07-09T06:17:39Z

tests/multi_net/asyncio_tcp_server_client_ipv6.py

+
+try:
+    # Check if IPv6 is supported
+    socket.socket(socket.AF_INET6, socket.SOCK_STREAM)


This detection won't work. All ports based on bare-metal lwIP will let this pass even if they don't have IPv6 enabled.

I don't know if it's possible to detect IPv6??

dpgeorge · 2025-07-09T06:18:01Z

tests/multi_net/asyncio_tcp_server_client_ipv6.py

+    print("SKIP")
+    raise SystemExit
+
+PORT = 8001  # Different from other tests to avoid conflicts


I don't think this needs to use a different port, 8000 should be fine.

dpgeorge · 2025-07-09T06:19:08Z

tests/multi_net/asyncio_tcp_server_client_ipv6.py

+
+async def handle_connection(reader, writer):
+    # Test that peername exists
+    peer = writer.get_extra_info("peername")


The peer variable is unused. I don't think you need to call this function at all.

dpgeorge · 2025-07-09T06:23:45Z

tests/multi_net/asyncio_tcp_server_client_ipv6.py

+
+    data = await reader.read(100)
+    print("read:", data)
+    assert data == message, "Data mismatch"


Don't need this assert, the .exp will test the data is correct.

dpgeorge · 2025-07-09T06:29:25Z

tests/net_hosted/asyncio_start_server_ipv6.py

+
+try:
+    # Check if IPv6 is supported
+    socket.socket(socket.AF_INET6, socket.SOCK_STREAM)


As above, this detection won't work.

dpgeorge · 2025-07-09T06:30:22Z

tests/net_hosted/asyncio_start_server_ipv6.py

+    try:
+        # Connect with IPv6 client
+        print("connect to ipv6 server")
+        reader, writer = await asyncio.open_connection("::", PORT)


For this to work requires loopback mode in the TCP/IP driver. That's fine on unix, but it'll fail on most bare-metal ports.

dpgeorge · 2025-07-09T06:30:36Z

tests/net_hosted/asyncio_start_server_ipv6.py

+        # Read response
+        data = await reader.read(100)
+        print("read:", data)
+        assert data == test_msg, "Data mismatch"


Assert not needed.

dpgeorge · 2025-07-09T06:35:00Z

tests/run-multitests.py

@@ -79,19 +79,59 @@ def globals(**gs):
            print("SET {{}} = {{!r}}".format(g, gs[g]))
        multitest.flush()
    @staticmethod
-    def get_network_ip():
+    def _get_ip_from_ifconfig(_nic, ipv6=False):
+        # Helper to get IP address from an interface object using appropriate format


This comment doesn't add much, and I'm confused what "appropriate format" means?

dpgeorge · 2025-07-09T06:35:06Z

tests/run-multitests.py

+            pass
+
+        # Find active network interface
+        try:


The code below is very hard to understand. Why so many try/except's?

Ensures that the underlying socket is opened with the correct protocol as parsed by getaddrinfo(). Signed-off-by: Andrew Leech <andrew.leech@planetinnovation.com.au>

andrewleech · 2025-07-31T06:26:20Z

The change to the multi test runner and new (duplicated) ipv6 tests have been removed from this PR to simplify it.

Updates to support testing ipv6 have been added in a different way in #17793

dpgeorge

Thanks for updating, looks good now.

dpgeorge · 2025-07-31T13:19:15Z

Actually, on second thoughts, I'm wondering if the family argument to start_server() is needed at all? It looks like that's not used in the tests in #17793, so maybe it's a useless (or rarely used) argument?

Adding that argument does add a fair amount to the code size. Whereas just passing through the family from getaddrinfo to the socket costs almost nothing.

So, if you want to get this in sooner rather than later, I suggest reverting the family argument addition.

andrewleech force-pushed the asyncio_server_ipv6 branch from 3443947 to 56e8e68 Compare May 16, 2025 01:26

dpgeorge added the extmod Relates to extmod/ directory in source label May 20, 2025

dpgeorge added this to the release-1.26.0 milestone May 20, 2025

andrewleech force-pushed the asyncio_server_ipv6 branch 2 times, most recently from 6227c48 to 3b17ecc Compare May 27, 2025 05:57

dpgeorge reviewed Jul 9, 2025

View reviewed changes

extmod/asyncio/stream: Add ipv6 support to start_server().

6684282

Ensures that the underlying socket is opened with the correct protocol as parsed by getaddrinfo(). Signed-off-by: Andrew Leech <andrew.leech@planetinnovation.com.au>

andrewleech force-pushed the asyncio_server_ipv6 branch from 3b17ecc to 6684282 Compare July 30, 2025 23:41

dpgeorge approved these changes Jul 31, 2025

View reviewed changes

dpgeorge modified the milestones: release-1.26.0, release-1.27.0 Aug 1, 2025

Uh oh!

extmod/asyncio/stream.py: Add ipv6 support to start_server(). #17311

Are you sure you want to change the base?

extmod/asyncio/stream.py: Add ipv6 support to start_server(). #17311

Conversation

andrewleech commented May 16, 2025

Summary

Testing

Trade-offs and Alternatives

Uh oh!

codecov bot commented May 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

github-actions bot commented May 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

andrewleech commented Jul 31, 2025

Uh oh!

dpgeorge left a comment

Choose a reason for hiding this comment

Uh oh!

dpgeorge commented Jul 31, 2025

Uh oh!

Uh oh!

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

codecov bot commented May 16, 2025 •

edited

Loading

github-actions bot commented May 16, 2025 •

edited

Loading