Skip to content

Lesson: ApolloExplorer; pitfall: v4net.device full-size frames; V4 boot time - #55

Merged
WDrijver merged 2 commits into
mainfrom
lesson/2026-11-apolloexplorer
Nov 16, 2026
Merged

WDrijver merged 2 commits into
mainfrom
lesson/2026-11-apolloexplorer

Conversation

@WDrijver

@WDrijver WDrijver commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

Summary

From the ApolloExplorer 1.4.4 work (WDrijver/ApolloExplorer#9), tested under ApolloOS and AmigaOS on the same V4:

  • New: os/pitfalls/v4net-full-size-frames-lost.md. v4net.device 3.0-beta never transmits a 1514-byte Ethernet frame (a 1500-byte IP packet); 1512 bytes and smaller work.
    • It still reports MTU 1500, so any TCP connection from the V4 stalls as soon as it sends more than ~1.4 KB.
    • The same bug shows under ApolloOS R9.6 (AROSTCP) and AmigaOS 3.9 (AmiTCP_NG 4.1.8): it is the driver.
    • Workarounds per stack:
      • AmiTCP_NG: mtu=1492 in DEVS:NetInterfaces/V4net, which covers every program. AmiTCP_NG does not implement TCP_MAXSEG.
      • AROSTCP: TCP_MAXSEG on the connected socket. It can only lower the negotiated MSS.
    • The fix belongs in the driver. Deprecate this file once a fixed version is out.
  • New: lessons/2026-11-apolloexplorer.md. What the project learned:
    • the "unstable network" was the driver
    • how TCP_MAXSEG, socket timeouts and SystemTagList() priority behave in AROSTCP, AmiTCP_NG and ApolloOS DOS. SystemTagList() ignores an NP_Priority tag.
    • remote configuration when command output can't come back
    • what the AI got wrong along the way
  • Changed: guides/testing-on-v4-hardware.md.
    • The V4 answered ping 34.4 s after power-on under AmigaOS 3.9 (15 s to ash under ApolloOS R9.5).
    • Rule: poll every second and give up after 60 s.
    • Also notes that nothing answers remotely unless ApolloExplorerSrv starts at boot.
  • Changed: tools/v4/v4test.py power-cycle. It now polls every second for at most 60 s (was every 5 s for 120 s).

Evidence

  • Measured on a V4 UniCorn, Core 12856, v4net.device 3.0, 2026-11-16:
    • ApolloOS R9.6 and AmigaOS 3.9 (Kickstart 45.57, AmiTCP_NG 4.1.8)
    • ping size sweeps and the protocol test harness tests/aetest.py in WDrijver/ApolloExplorer
  • Source:
    • AROSTCP tcp_usrreq.c, tcp_output.c, kern/uipc_socket.c
    • AmiTCP_NG src/netinet/tcp_usrreq.c, src/net/ng_ifconfig.c
    • ApolloOS rom/dos/systemtaglist.c, createnewproc.c

Notes for review

  • applies_to.os: [apolloos-r10, amigaos-3.9], but the ApolloOS measurements were on R9.6. The body and sources say so. AmigaOS 3.2 is untested.
  • Workaround in a pitfall: the pitfall documents a workaround for a driver bug. v4net.device is from the apollo-core SVN, not ApolloOS, and the bug is the same under AmigaOS, so the rule against ApolloOS-bug workarounds does not apply.
  • Not in the pitfall:
    • Occasional packet loss (one per ~4 MB) seen only under ApolloOS. Under AmigaOS with the same driver there was none. It stays an open point in the lesson.
    • SystemTagList()/NP_Priority stays a lesson candidate until it is checked on AmigaOS 3.x.

Checks

check_frontmatter.py, check_links.py, check_owners.py, check_saga_addresses.py and markdownlint all pass locally; v4test.py compiles.

🤖 Generated with Claude Code

lessons/2026-11-apolloexplorer.md: what the ApolloExplorer 1.4.4 work
learned.  An "unstable network" was the V4 Ethernet driver; the
AROSTCP and DOS sources answered how TCP_MAXSEG, socket timeouts and
SystemTagList() priority behave; and what the AI got wrong on the way.

os/pitfalls/v4net-full-size-frames-lost.md: v4net.device 3.0-beta
never transmits a 1514-byte frame (1500-byte IP packet), which stalls
any TCP connection that sends more than ~1.4 KB.  Measured on a V4
UniCorn, Core 12856, ApolloOS R9.6; workaround with TCP_MAXSEG on the
connected socket until the driver is fixed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@WDrijver
WDrijver requested a review from TuKo1982 as a code owner November 16, 2026 12:44
The same v4net.device 3.0 bug shows under AmigaOS 3.9 with AmiTCP_NG
4.1.8, so it is the driver, not ApolloOS.  AmiTCP_NG does not implement
TCP_MAXSEG (EINVAL, "not yet"), so the per-socket workaround fails
silently there; mtu=1492 in DEVS:NetInterfaces/V4net fixes it for
every program (measured: all ApolloExplorer tests pass, 4.2 MB/s).

The occasional packet loss seen under ApolloOS is removed from the
pitfall: under AmigaOS 3.9 with the same driver and core there was
none, so it is not the driver alone.  The lesson keeps it as an open
point and adds what the AI got wrong about it and about TCP_MAXSEG.

guides/testing-on-v4-hardware.md: the V4 answered ping 34.4 s after
power-on under AmigaOS 3.9 (15 s to ash under ApolloOS R9.5): poll
every second and give up after 60 s, and note that nothing answers
remotely unless ApolloExplorerSrv starts at boot.  v4test.py
power-cycle now polls every second for at most 60 s (was 5 s / 120 s).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@WDrijver WDrijver changed the title Lesson: ApolloExplorer, and pitfall: v4net.device full-size frames Lesson: ApolloExplorer; pitfall: v4net.device full-size frames; V4 boot time Nov 16, 2026
@WDrijver
WDrijver merged commit 66fa4db into main Nov 16, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant