rt: netconsole (dynamic, configfs) and sysrq in the 7.2 kernel fragment - #140
Merged
Conversation
…nt — the board's only way to say why it rebooted A whole-board reboot during a PSX session on 2026-09-02 left no trace: the serial console is disabled, pstore has no RAM reservation, and /tmp died with the reboot. NETCONSOLE=y with the DYNAMIC target so the listener address stays in the rig's out-of-band config, never in this tree; MAGIC_SYSRQ for a live dump while the board is wedged but reachable. RT/7.2 fragment only; the 6.18 stock-parity kernel is untouched. Leaves with the other debug options when a shipping 7.2 kernel is cut. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012AZB1JmCKWaJt1ka22oex3
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
The SysRq default enable mask as set (0x1) enables all SysRq functionality and the fragment also duplicates symbols already inherited from the shared linux.config, both of which should be corrected/clarified before merge.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Pull request overview
Adds RT/7.2-only kernel troubleshooting facilities (netconsole via configfs and SysRq defaults) in the RT kernel config fragment to improve post-mortem/debug visibility when the board reboots or wedges and local logs are lost.
Changes:
- Enable netconsole with dynamic (configfs) runtime target configuration in the RT/7.2 kernel fragment.
- Enable SysRq and set a default SysRq enable mask in the RT/7.2 kernel fragment.
- Add explanatory comments documenting the intended troubleshooting workflow and motivation.
File summaries
| File | Description |
|---|---|
| board/mister/de10nano/linux-rt.fragment | Enables netconsole (dynamic/configfs) and SysRq defaults for RT/7.2 troubleshooting output. |
Review details
- Files reviewed: 1/1 changed files
- Comments generated: 2
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Copilot review on #140: CONFIGFS_FS=y and MAGIC_SYSRQ=y already come from the shared linux.config, and MAGIC_SYSRQ_DEFAULT_ENABLE=0x1 is the kernel default the stock-parity 6.18 kernel also has; restating them in the RT fragment blurred what is RT-only. Dropped; the comment now says they are inherited. Deliberately NOT narrowing the sysrq mask to 0x8: 0x1 (= all functions) is what stock ships, and diverging RT from stock on a debug knob buys nothing. Merging old vs new fragment onto the 7.2.3 RT .config yields the same symbol values. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYjrnzEEu4HzJSPWNAaRvm
Contributor
There was a problem hiding this comment.
🟡 Changes recommended
The comments contradict the existing ramoops reservation, and several claimed RT-only settings are shared-config no-ops.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Review details
Suppressed comments (1)
board/mister/de10nano/linux-rt.fragment:165
- These are not RT-fragment deltas: the shared
linux.configalready setsCONFIG_CONFIGFS_FSandCONFIG_MAGIC_SYSRQ(lines 525 and 565), and Linux 7.2 already defaultsCONFIG_MAGIC_SYSRQ_DEFAULT_ENABLEto0x1. Thus these three lines are no-ops and make the block's “RT/7.2 variant ONLY” claim misleading, contrary to this fragment's stated only-deltas convention. Keep only the netconsole deltas here; SysRq remains available through the shared config.
# stock-parity 6.18 kernel -- narrowing it here
# would make RT diverge from stock for no gain.
# Measured need: a whole-board reboot during a PSX session on 2026-09-02 left no
- Files reviewed: 1/1 changed files
- Comments generated: 1
- Review effort level: Balanced
…visioned Copilot review on #140: the comment claimed the pstore backend had no RAM reservation, but the RT series carries 0046-dts-ramoops.patch (1 MiB at 0x1FE00000) and PSTORE_RAM=y. Reworded: ramoops is the on-board post-mortem channel that needs a successful warm reboot to read back; netconsole streams printk off the board live, which is what a hang or power cycle leaves you. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FYjrnzEEu4HzJSPWNAaRvm
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Troubleshooting channels for the RT/7.2 variant only; the 6.18 stock-parity kernel is untouched. Leaves with the other debug options when a shipping 7.2 kernel is cut.
Why now. A whole-board reboot during a PSX session on 2026-09-02 left no trace anywhere: the serial console is disabled, the pstore backend the fragment already provisions has no RAM reservation in the device tree, and
/tmpdied with the reboot (kernel.panic=15fired). Netconsole is the only path that gets kernel text off this board.What changes in
board/mister/de10nano/linux-rt.fragment:CONFIG_NETCONSOLE=y,CONFIG_NETCONSOLE_DYNAMIC=y(+CONFIGFS_FS) so the target is set at RUNTIME through/sys/kernel/config/netconsolefrom the rig's out-of-band config — no listener address is baked into this tree;CONFIG_MAGIC_SYSRQ=ywith a default mask of 1 for a live backtrace/state dump while a wedged board is still reachable.Verified:
make rtrc=0 on the 7.2.3 pin; the built.configcarries all five symbols;zImage_dtb9,305,438 B and a 3.6 MB7.2.3module tree produced. The kernel is staged on the DE10 for its first boot with netconsole armed toward the LAN collector.🤖 Generated with Claude Code
https://claude.ai/code/session_012AZB1JmCKWaJt1ka22oex3