Skip to content

fix(boot): skip the boot screen when no panel responds - #179

Merged
jonasniesner merged 1 commit into
OpenDisplay:mainfrom
nlimper:fast-boot-without-panel
Nov 4, 2026
Merged

jonasniesner merged 1 commit into
OpenDisplay:mainfrom
nlimper:fast-boot-without-panel

Conversation

@nlimper

@nlimper nlimper commented Nov 1, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Booting with no panel connected (or a panel that does not respond) took about 10 minutes: every BUSY wait in the boot-screen path ran into its timeout. This adds a short presence check before the boot screen.

bootPanelResponds() pulses the panel reset and waits up to 2 s for BUSY to report idle (HIGH on UC81xx, LOW on SSD16xx). After a hardware reset every controller does that within milliseconds, which is also what the GoodDisplay sample code waits for. If BUSY stays busy, initDisplay() logs a warning, cuts the rail with pwrmgm(false) and skips the boot screen.

  • Boot without a panel: ~10 min → ~4 s.
  • pwrmgm(false) instead of epdSessionForceOff(): ForceOff sends sleep commands and waits on BUSY, which is exactly what hangs here (same pattern as the FastEPD init-failed tail).
  • Dual chip-select (BBEP_SPLIT_BUFFER) panels, and configs without a BUSY or RESET pin, are not probed (untested), so their behaviour is unchanged.
  • No change to image uploads after boot.

Testing

On a Seeed EN04 (nRF52840), with this change on top of a branch that also adds new panels (not part of this PR):

  • No panel connected: boot completes in ~4 s, log shows Panel not responding (BUSY stays busy after reset) - skipping boot screen, BLE works and a config can be written.
  • Panels connected: boot screen as before for the GDEY0266F52H (57), GDEM0122T61 (13, SSD1681) and GDEY029F51H (30).

Builds against the current bb_epaper pin: nrf52840custom, esp32-s3-N16R8, esp32-c3-N4, esp32-wrover-e-N4R8.

🤖 Generated with Claude Code

After a reset every controller reports idle on BUSY within milliseconds.
If BUSY stays busy for 2 s there is no panel (or a dead one), and every
later BUSY wait would run into its timeout: booting without a panel took
about 10 minutes, now about 4 seconds. Dual chip-select panels are not
probed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@nlimper
nlimper requested a review from jonasniesner as a code owner November 1, 2026 15:31
@jonasniesner
jonasniesner merged commit f55c6d8 into OpenDisplay:main Nov 4, 2026
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.

2 participants