[v0.20.x-branch] Backport #10540: discovery: fix gossiper shutdown deadlock

[github-actions]

Backport of #10540

---

When processing a remote network announcement, it is possible for two error messages to be sent back on the errChan. Since Brontide doesn't actually read from errChan, and since errChan only buffered one error message, the sending goroutine would deadlock forever. This would only become apparent when the gossiper attempted to shut down and got hung up.

For now, we can fix this simply by buffering up to two error messages on errChan. There is an existing TODO to restructure this logic entirely to use the actor model, and we can do a more thorough fix as part of that work.

This bug was discovered while doing full node fuzz testing and was triggered by sending a specific channel_announcement message and then shutting down LND.

[github-actions]

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin backport-10540-to-v0.20.x-branch
git worktree add --checkout .worktree/backport-10540-to-v0.20.x-branch backport-10540-to-v0.20.x-branch
cd .worktree/backport-10540-to-v0.20.x-branch
git reset --hard HEAD^
git cherry-pick -x 5fa025e9e7a180de0a81750757fcc20eb37b28a9
git push --force-with-lease

[lightninglabs-deploy]

🟠 PR Severity: HIGH

Backport bugfix | 1 functional file | 9 lines changed

🟠 High (1 file)

• discovery/gossiper.go - Gossip protocol shutdown deadlock fix

🟡 Medium (0 files)

None

🟢 Low (1 file)

• docs/release-notes/release-notes-0.20.1.md - Release notes update

Analysis

This PR backports #10540 which fixes a shutdown deadlock in the gossiper. The primary change is in discovery/gossiper.go, which is part of the gossip protocol subsystem classified as HIGH severity.

Key observations:

• Focused bugfix: Only 9 lines changed in the core functional file
• Shutdown safety: Addresses a deadlock condition during node shutdown
• Test coverage: Includes comprehensive test case to verify the fix
• Backport: This is a backport to the v0.20.x branch, indicating the fix has already been reviewed in master

The HIGH severity is appropriate because the gossiper is responsible for network topology propagation, and while this is a shutdown-specific bug rather than affecting normal operation, changes to concurrency control in the discovery package warrant careful review by an engineer familiar with the gossip protocol's threading model.

---

_{To override, add a severity-override-{critical,high,medium,low} label.}

[Roasbeef]

Fixed the conflict manually.

[gijswijs]

Confirmed that the backport is the same fix that's already in master, so:
LGTM! 🎉