6.5.7 Firmware for Prusa CORE One L  

RE: 6.5.7 Firmware for Prusa CORE One L

Thanks but I think I'll wait a week or two before updating my printer. 

RE: 6.5.7 Firmware for Prusa CORE One L

Not in the release notes. What is interesting is the commit comments, Prusa most definitely used a LLM (looks like Claude) for this release. 

Crash / stability fixes (ungated core code — high value, invisible in notes)

• 81d49d4bf — logging deadlock guard (BFW-8838). Task::send() now drops the event if interrupts are globally masked (__get_PRIMASK() != 0). Previously, logging with interrupts disabled blocked forever waiting on a semaphore the (unschedulable) log task could never release. Hard-deadlock fix.
• 320f0aa41 — Power-Panic semaphore crash (BFW-8158). binary_semaphore acquire()/release_blocking() retried instead of std::abort() when a portMAX_DELAY wait returns early. Power Panic calls xTaskAbortDelay(), which can wake those "wait forever" calls — the old code treated that as impossible and intentionally crashed. Removes a Power-Panic crash path.
• a26c32dea — ISR stack 1640→2048 B (master boards, incl. Core One's xBuddy). Headroom against MSP-stack overflow / corruption (BFW-8196/8577).
• 34867d3bf — abort SPI DMA before deinit (BFW-8212). HAL_DMA_Abort() before HAL_DMA_DeInit() so a busy transfer doesn't leave the peripheral wedged. Applies to xBuddy SPI buses.
• 3eb8777f7 — Connect on the error-boot path. Split the connect task dependency so Connect can come up while an error screen is shown (the error path never signals autostart_done, which previously hung Connect startup).
• 5a49103fd — stop logging coordinate reports (report_current_position()) to the printer log — log-hygiene, BFW-8517.

User-visible behavior changes (undocumented)

• e78d567bc — firmware now disables E + XY steppers at end of print. pre_finalize_print() calls disable_e_steppers(); disable_xy_steppers(); unconditionally (was the slicer's End-G-code job). Side effect flagged in the commit itself: XY homed position is lost after a completed print — next move needs a re-home. Worth knowing if you rely on End-G-code behavior.
• Abort/park safety trio (BFW-8179, all ungated → Core One):
  • b3a95b07b — on print abort, X/Y are marked unhomed (position unreliable after quick_stop()), so the head only Z-lifts and no longer does an XY park move.
  • 3503df5da — mapi::park() becomes a no-op when X/Y are unhomed (the enforcement that makes the above safe).
  • 69e3d5013 — retract + Z-lift always happen on abort even when the XY park is skipped.
• 3e13458bc — unload skips the park move when filament is already safely auto-retracted (Core One has HAS_AUTO_RETRACT).
• 7ef1a641d — auto-retract re-heats the nozzle first (blocking M109 to filament temp, then restores prior target) so a cooled nozzle doesn't jam during retraction.
• ec58b03d — direct heatbreak-fan control (BFW-8504). The heatbreak (hotend-cooling) fan PID output is written straight to the fan object, bypassing the planner/FAN1-pin path, so it responds immediately instead of being subject to planner buffering/draining. Core One takes this via its PIDTEMPHEATBREAK single-hotend path. ⚠️ Caveat: this is labeled a "stopgap" and was reverted on master after the release — but within the v6.5.3→v6.5.7 range it is the shipped behavior.
• Status-LED changes (Core One has RGB status LEDs; switch logic is shared, not iX-gated):
  • a1056788d — LEDs pulse yellow whenever a Warning FSM is active.
  • 3802c298f — LEDs stay in the Printing animation through end-of-print park/unload, then go green only at true Finish (previously switched to green earlier).
  • 54d220cad — pure enum rename (Finishing→Finished, Aborting→Aborted), no behavior change.
• ab9b5b45b — power-panic awaiting-resume reported as Paused, not Attention (Connect/API/UI status). More accurate for an auto-resume.
• 993fcdec4 — Connect waits for the autostart/power-panic-resume sequence before connecting → no transient false "IDLE" reported to Prusa Connect mid-print after a power-loss reboot.
• 5ad3b3503 — assisted-load timeout 40 s → 120 s. ⚠️ Titled "iX" but the constant in pause.cpp is not printer-gated — it applies to all printers including Core One (likely an unintended scope/title mismatch).

RE:

I asked claude to summarize the above:

Prusa Core One — what actually changed from 6.5.3 to 6.5.7

The official release notes only list five things, but there's quite a bit more under the hood. 

Not in Release Notes:

• Fewer random freezes and crashes — a couple of rare lock-ups and power-loss-related crashes were quietly fixed, plus some general stability hardening.
• Better power-loss recovery — after a power blip the printer now correctly shows as "paused/resuming" instead of throwing an attention error, and it no longer briefly flashes "idle" to Prusa Connect mid-print.
• Smarter status lights — they pulse yellow when there's a warning, and stay in the "printing" color through the final park/unload instead of going green too early.
• Nozzle stays hot during retraction — avoids jams when the nozzle had cooled too much.
• Safer print cancels — after you stop a print, the printer no longer makes risky moves based on a position it can't trust.
• More patience loading filament — the "insert filament" timeout was extended, so it gives you more time before giving up.

Heads-up / things to know:

• The printer now relaxes its motors at the end of a print and "forgets" its exact position — totally normal, but it'll re-home before the next move.
• One hotend-fan tweak in this version was a temporary fix that's already been reversed in later development builds — nothing to do, just don't be surprised if it changes again in a future release.