Skip to content

troubleshooting.md

Latest commit

 

History

History
52 lines (43 loc) · 2.97 KB

troubleshooting.md

File metadata and controls

52 lines (43 loc) · 2.97 KB
Raw

Troubleshooting Guide

This guide addresses common issues encountered when using LabWired Core.

Simulation Issues

"Memory Violation at 0x..."

Symptom: The simulator halts with StopReason::MemoryViolation(address). Cause: The firmware attempted to access an address that is not mapped to Flash, RAM, or any registered peripheral. Solution:

  1. Check your chips/<chip>.yaml definition. Does the address fall within the declared Flash or RAM regions?
  2. Check your system.yaml. Is there a peripheral missing that should be at that address?
  3. Verify the firmware linker script (memory.x). It might be placing data in valid hardware regions that the simulator doesn't know about yet.

"Instruction Decode Error"

Symptom: StopReason::DecodeError(address). Cause: The CPU attempted to execute an instruction that is invalid or undefined for the current architecture (e.g., executing data as code). Solution:

  1. Check the Vector Table. Is the Reset Vector pointing to the correct entry point (usually _start or Reset_Handler + 1 for Thumb mode)?
  2. Verify load_firmware succeeded. Run with --trace to see the first few instructions. If it crashes immediately, the entry point might be wrong.
  3. Cortex-M requires the PC LSB to be 1 (Thumb mode). If your vector table has an even address (e.g., 0x08000100), the CPU will switch to ARM mode and fault.

"DAP Server Unreachable"

Symptom: VS Code displays "Connection refused" or "Timeout" when starting a debug session. Cause: The LabWired instance isn't running, or the port is blocked. Solution:

  1. Ensure LabWired is started with --gdb or is in interactive mode.
  2. Check the port (default 3333). Is it in use by another OpenOCD/J-Link instance?
  3. In launch.json, verify miDebuggerServerAddress matches localhost:<port>.

"UART Output is Garbage"

Symptom: uart.log contains random characters. Cause: Baud rate mismatch between firmware and simulator config. Solution:

  1. LabWired's UART model is currently "byte-perfect" (it doesn't simulate physical baud timing errors), but ensure your firmware is actually writing valid ASCII.
  2. Check if echo_stdout is enabled or if you are looking at raw binary data.

CI/CD Issues

"Max Steps Reached"

Symptom: CI fails with StopReason::MaxSteps. Cause: The proprietary max_steps limit in your test script is too low for the initialization code to complete. Solution:

  1. Increase max_steps in your .yaml test script.
  2. Optimize your firmware boot sequence (e.g., reduce PLL validation timeouts).

"No Progress / Stuck"

Symptom: StopReason::NoProgress. Cause: The PC has stayed in the same range (or exactly the same address) for no_progress_steps. This usually means an infinite loop (e.g., while(1) {} or a HardFault handler). Solution:

  1. Use --trace to identify the loop.
  2. If it's a valid strict polling loop, insert a asm!("nop") or similar to vary the PC, or increase no_progress_steps.