arch/sim: Stop publishing stale X11 display during teardown.

[Zepp-Hanzj]

Summary

• Fixes a shutdown ordering issue in the sim X11 framebuffer teardown path.
• sim_x11events() may continue polling g_display from the idle loop while the X11 connection is being torn down.
• Save the current Display * in a local variable, clear the global g_display before teardown, and use the local pointer for the remaining X11 cleanup calls.
• Related NuttX Issue: none.
• Related NuttX Apps Issue/Pull Request: none.

Impact

• Is new feature added? NO.
• Is existing feature changed? YES: only the sim X11 framebuffer shutdown ordering is changed.
• Impact on user? NO expected user-visible behavior change; the change avoids publishing a stale Display pointer during shutdown.
• Impact on build? NO.
• Impact on hardware? NO real hardware impact; this is limited to the simulator X11 framebuffer path.
• Impact on documentation? NO.
• Impact on security? NO.
• Impact on compatibility? NO API, ABI, or configuration compatibility change.
• Dependencies: none.

Testing

I confirm that changes are verified on local setup and work as intended:

• Build Host(s): Linux DESKTOP-6B1686O 6.6.114.1-microsoft-standard-WSL2 x86_64, gcc (Ubuntu 10.5.0-1ubuntu1~20.04) 10.5.0.
• Target(s): sim:nsh.
• Static/style check: ./tools/checkpatch.sh -g HEAD~...HEAD.
• Build test: ./tools/configure.sh sim:nsh && make -j4.
• Runtime smoke test: booted ./nuttx, ran help, uname -a, and poweroff.

Testing logs before change:

Code inspection: before this change, sim_x11uninit() kept g_display published while
the teardown path called XShmDetach(), XUngrabButton(), and XCloseDisplay() through
g_display. sim_x11events() also polls g_display from the idle loop, so the event
path could observe a stale Display pointer during shutdown.

Testing logs after change:

$ ./tools/checkpatch.sh -g HEAD~...HEAD
✔️ All checks pass.

$ ./tools/configure.sh sim:nsh
# configuration written to .config

$ make -j4
LD:  nuttx
Pac SIM with dynamic libs..
SIM elf with dynamic libs archive in nuttx.tgz

$ printf 'help\nuname -a\npoweroff\n' | timeout 10s ./nuttx
NuttShell (NSH) NuttX-12.13.0
nsh> help
help usage:  help [-v] [<cmd>]
...
nsh> uname -a
NuttX 12.13.0 a1a048d9f6 May 25 2026 23:55:29 sim sim
nsh> poweroff

PR verification Self-Check

• This PR introduces only one functional change.
• I have updated all required description fields above.
• My PR adheres to Contributing Guidelines and Documentation (git commit title and message, coding standard, etc).
• My PR is still work in progress (not ready for review).
• My PR is ready for review and can be safely merged into a codebase.

[jerpelea]

please follow the contribution guidelines and add a proper PR title, description and test cases

[Zepp-Hanzj]

Addressed the requested PR metadata/test details:

• Amended the commit title/message to follow the contribution guideline format.
• Updated the PR title and description with Summary, Impact, Testing, and Self-Check sections.
• Added local validation details for ./tools/checkpatch.sh -g HEAD~...HEAD, ./tools/configure.sh sim:nsh && make -j4, and a ./nuttx runtime smoke test.

CI has been re-triggered on the updated commit.

[Zepp-Hanzj]

@jerpelea The requested PR metadata/test details have been addressed in the current head commit:

• PR title now follows the NuttX contribution guideline format.
• PR description includes Summary, Impact, Testing, and Self-Check sections.
• Testing details include checkpatch, sim:nsh build, and runtime smoke-test evidence.
• Current CI is green.

Could you please re-review when convenient so the stale change-request state can be cleared if this now looks OK? Thanks.

[Zepp-Hanzj]

🎉 Congratulations on the merge! Thank you for the contribution — glad to see this sim/X11 teardown fix landed successfully. Much appreciated!