Stable identifiers produced by the core (SPEC-001 §7). Each has a severity default and a subject.
Every diagnostic in a report carries a phase: load for what loading the
project found (netlist checks, missing images, probes), run for what the
simulation raised. A run's reset forgets the run's own diagnostics and keeps
the load's, so embsim run --json, scenario reports and diagnostics.forbid
see both; loading another project forgets them all. embsim validate reports
the load's and those of the circuit settled at its first instant (firmware not
run).
Each message is in English; in JSON a record also carries messages.fr, the
same message in French (SPEC-001 §7), which the workbench shows when its
language is French.
| Code | Default severity | Subject | Meaning |
|---|---|---|---|
schema.parse |
error | file | a description is not valid JSON |
schema.unknown_field |
error | file, node | a description contains a field its schema does not define |
schema.missing_field |
error | file, node | a required field is absent |
schema.type |
error | file, node | a field has the wrong type or an out-of-range value |
schema.version |
error | file | unsupported schema_version |
catalog.not_found |
error | id | a referenced component, board or MCU is not in any library |
catalog.pin_unknown |
error | instance, pin | a net references a pin the component does not have |
catalog.pin_twice |
error | instance, pin | a pin appears in two nets |
catalog.composite_depth |
error | id | composites nest deeper than 8 levels |
catalog.model_unsupported |
error | id | a composite is used where only a primitive can be |
catalog.port_missing |
warning | id, pin | a composite has no port for one of its pins |
catalog.instantiate |
error | id | a component could not be instantiated |
fidelity.tier_missing |
error | id | the session policy needs a fidelity tier the model does not declare |
analog.param_out_of_range |
error | instance, param | a parameter is outside its declared range |
net.contention |
warning | net | two drivers of equal strength disagree (two opposing weak pulls are a divider and are not reported) |
net.floating_input |
warning | pin | an input sampled a net with no driver and no pull |
net.unknown_input |
warning | pin | an input sampled X |
net.power_short |
error | net | two supplies with different nominal voltage on one net |
net.unconnected_required_pin |
warning | pin | a required pin is unconnected |
sched.delta_overflow |
fatal | targets | zero-delay propagation exceeded max_delta_cycles |
sched.time_overflow |
error | component | a delay overflows the time horizon |
sched.queue_full |
fatal | the event queue capacity was exceeded | |
analog.no_convergence |
error (fatal in strict) | nodes | the operating point or a step did not converge |
analog.step_failure |
fatal | island | consecutive step rejections exceeded the limit |
analog.timestep_too_small |
fatal | device | the step fell below h_min |
analog.singular_topology |
error | node | a floating node or inconsistent sources |
analog.implicit_supply |
info | pin | a pull or bridge used the family's nominal supply |
analog.implicit_ground |
info | node | an island with no ground reference was given one |
analog.supply_driven |
info | net | a power pin's net is driven by a component, so its ideal supply is not added |
analog.overvoltage, analog.overcurrent |
warning | device | a declared absolute maximum exceeded |
island.requires_electrical |
error | instance | a component has no functional model under functional policy |
mcu.no_firmware |
info (warning for an input) | board | a board has no firmware: its MCU's pins stay inputs; an input to it has nowhere to go |
mcu.no_image |
warning | board | the firmware's image file is missing: the core is held in reset until one is loaded |
mcu.stale_image |
warning | board | an ELF image is older, by more than a minute, than a source file its line tables name: it was built before that source last changed (embsim run --build rebuilds it first) |
mcu.irq_storm |
warning | component | one exception was entered again and again, more often than once per 400 core cycles over 10 ms: what raises it stays set (a flag its handler does not clear, a level-triggered input held active); the message names the interrupt and its peripheral's status registers as they read |
project.no_connect |
error | project | no_connect names a pin no part or board has |
project.no_connect_wired |
warning | pin | a pin no_connect lists is on a net after all |
perf.hot_net |
info | net | one net's edges explain most of the run's analog steps: each re-solves its island (embsim run --circuit-profile shows the rest) |
perf.island_size |
warning | net | a busy logic line is joined to a large analog island only through a resistor to a supply rail (a pull-up drawn to the board's 3V3): each edge solves the whole island; emblink/pull_up to the rail's voltage keeps the line on its own with the same levels |
mcu.image |
error | board | the image could not be loaded (not ELF or HEX, wrong machine, outside memory) |
mcu.register_file |
error | board | the MCU's register description is missing or invalid |
mcu.model_missing |
info | board, peripheral | a peripheral is bound to a model this build does not have: its registers still hold values |
mcu.unsupported_access |
warning | board, register, pc | the first access to a peripheral with registers and no model |
mcu.clock_gated |
warning | board, register, pc | an access to a peripheral whose clock is off: reads 0, writes ignored |
mcu.bus_error |
warning | board, address, pc | nothing answered an access: the core takes a bus fault |
mcu.held_in_reset |
warning | board, register, pc | an access to a peripheral its reset controller holds in reset (the RP2040's RESETS): reads its reset values, writes ignored |
mcu.usb_fifo |
warning | board | an STM32 OTG core's FIFOs as the program placed them: an IN endpoint's FIFO overlapping the receive FIFO or past the 320 words of FIFO RAM, or the receive FIFO read before GRXSTSP took a packet's status |
usb.dfu |
warning | component | a USB host part's DFU download stopped: the device's status, a request it refused, an image outside its memory layout or a read-only segment, a file it could not read |
usb.enumeration |
warning | component | a USB host part could not enumerate the device on its cable: the request that failed and why (a STALL, no answer, 5 s); its log has the steps |
mcu.usb_boot |
warning or info | board | the RP2040 goes to its USB boot (the program called reset_to_usb_boot, a warning; BOOTSEL held at reset, info): the RPI-RP2 drive on its USB cable |
mcu.uf2_written |
info | board | every block of a UF2 file copied onto the RP2040's USB boot drive is in flash: the part reboots into it |
mcu.no_boot2 |
warning | board | the RP2040's bootrom finds no boot2 with a good CRC in flash: the part waits in its USB boot (the RPI-RP2 drive) |
mcu.rescue |
info | board | the RP2040 was reset by its Rescue DP (a probe set the Rescue DP's CDBGPWRUPREQ, then cleared it): its bootrom halts in its dead loop for the debugger to take over |
mcu.watchdog_reset |
info | board | the watchdog reset the chip (the RP2040's counter reached 0 or CTRL.TRIGGER; the FE310's comparator fired with wdogrsten) |
mcu.low_power_reset |
info | board | an STM32F4 entered Stop or Standby while its option bytes' nRST_STOP or nRST_STDBY is 0: the part resets instead (LPWRRSTF) |
mcu.easydma_range |
warning | board, peripheral | an nRF52840 EasyDMA buffer outside Data RAM: nothing is moved |
mcu.locked_write |
warning | board, register, pc | a write to a PIC32's OSCCON while SYSKEY's unlock sequence has not been written, or to a byte of a PIC24's OSCCON without its own unlock sequence: ignored, as on the part (the first one) |
mcu.trap |
warning | board, pc | a PIC24 trap (an address, stack or math error, an oscillator failure): the core takes its vector; the first ten are reported |
mcu.boot_mode |
warning | board | BOOT0 selects system memory, which is not modelled |
mcu.stack_overflow |
error | board | with the stack analysis on (--stack, analysis.start({stack: true})), the main stack went past its limit (the link's __stack_limit, or into .data and .bss): the stack pointer, the limit and where it came from, the instruction and its function; once a reset, the run goes on |
mcu.fault |
error | board | with the fault report on (--faults, analysis.start({faults: true})), the core took a fault: what it was (a HardFault and why, the address), the instruction and its function and line, the handler; the whole report in analysis.faults |
analysis.unavailable |
warning | board | coverage or the profile was on, and a reloaded project's board cannot be measured (no firmware, no function symbols) |
mcu.exit |
info | board | the firmware exited through semihosting, with its status |
mcu.lockup |
error | board | the core locked up (a fault with nowhere to go) |
input.rejected |
warning | component | a component refused an input |
probe.invalid |
error | probe | a probe's target has an unknown kind (with a suggestion) or names nothing on the bench: the project does not load |
project.image_missing |
error | board | embsim run: a board's firmware image cannot be read; the run stops at load (exit 2) unless --allow-no-image |
io.read |
error | file | a file could not be read |
ext.budget_exceeded |
error | instance | a Lua evaluation exceeded its instruction budget |
limit.exceeded |
fatal | limit | a configured capacity was exceeded |