Adds the three firmware host-test modules required by the Testing Strategy as a
plain gcc harness under firmware/test/ — NOT idf.py --target linux. That path was
rejected (docs/notes/firmware-host-test-approach.md, bf-21t): firmware/main
cannot host-link because csi.c pulls in esp_wifi.h and provision.c pulls in
driver/uart.h, and the single `main` component REQUIRES esp_wifi/bt/driver,
which have no linux build — so even nvs_migration.c (hostable in isolation) is
unhostable as part of the component. The harness therefore tests dependency-free
logic extractions and binary-format/wire contracts instead of linking the
firmware source.
- test_nvs_migration.c: fresh-install init to v1, no-downgrade guard, forward
migration loop dispatch (v→v+1 at index v−1), and the concrete v1→v2 step
(rename ms_ip→mothership_ip, default ntp_server), driven against an in-memory
NVS store. Mirrors nvs_migration.c decision-for-decision.
- test_csi_frame.c: 24-byte header field round-trip, explicit little-endian
timestamp byte order, signed-RSSI (uint8_t) reinterpretation, I/Q payload
copy, n_sub=0 header-only probe, and the ingestion-side validation rules
(too-short / payload-mismatch / n_sub>128 / bad channel). Mirrors the
websocket.c encoder contract (offset/byte for offset/byte).
- test_serial_prov.c: provisioning JSON parser + NVS-mapping mirror of
provision.c (all four protocol branches + every field mapping), shipping a
bounded recursive-descent JSON decoder as the fuzz target. The fuzz pass
(4000 random byte streams, a tricky-input corpus, 500 deep-nesting cases)
proves the parser never crashes and the protocol always answers a single
well-formed {"ok":...} line on any UART input.
- Makefile: gcc build/run recipe that globs every test_*.c + test_runner.c.
CI wiring: the Dockerfile firmware-builder stage now runs `make -C test test`
before the expensive ESP-IDF build, so a logic/format-contract regression fails
the image build fast. .gitignore + .dockerignore exclude the regenerable
host_tests binary.
docs/plan/plan.md Testing Strategy updated from the idf.py description to the
gcc harness (matching the decision record).
28 tests, all passing. go test ./... and go vet ./... unchanged (firmware-only).
Co-Authored-By: Claude <noreply@anthropic.com>
123 lines
4.9 KiB
Docker
123 lines
4.9 KiB
Docker
# Spaxel Mothership Dockerfile
|
|
# Multi-stage build: ESP32 firmware (amd64 only) → Go binary → minimal runtime image
|
|
# Build arguments for multi-platform support
|
|
ARG TARGETPLATFORM=linux/amd64
|
|
ARG TARGETARCH=amd64
|
|
|
|
# Stage 1: Build ESP32-S3 firmware (amd64 only - ESP-IDF is x86_64)
|
|
FROM espressif/idf:v5.2 AS firmware-builder
|
|
ARG TARGETPLATFORM
|
|
|
|
# Create build directory
|
|
RUN mkdir -p /project/build
|
|
|
|
# Handle amd64-only firmware build: skip on arm64, build on amd64
|
|
RUN if [ "$TARGETPLATFORM" != "linux/amd64" ]; then \
|
|
echo "# Firmware not available on $TARGETPLATFORM (ESP-IDF is amd64-only)" > /project/build/spaxel-firmware-merged.bin && \
|
|
echo "Firmware build skipped - placeholder created"; \
|
|
fi
|
|
|
|
# Only copy firmware source and build on amd64 (placeholder already created on arm64)
|
|
RUN if [ "$TARGETPLATFORM" = "linux/amd64" ]; then \
|
|
cd /project && \
|
|
echo "Building ESP32 firmware for $TARGETPLATFORM"; \
|
|
else \
|
|
exit 0; \
|
|
fi
|
|
|
|
# Bust Kaniko layer cache when flash size config changes (sdkconfig.defaults).
|
|
# Without this ARG, Kaniko can serve a cached firmware layer that was built with
|
|
# the old 16MB config even after sdkconfig.defaults is updated to 4MB.
|
|
ARG FIRMWARE_CACHE_BUST=2026-06-03
|
|
|
|
WORKDIR /project
|
|
COPY firmware/ ./
|
|
|
|
# Remove any stale generated sdkconfig so set-target regenerates it from
|
|
# sdkconfig.defaults (which specifies CONFIG_ESPTOOLPY_FLASHSIZE_4MB=y).
|
|
RUN rm -f sdkconfig sdkconfig.old
|
|
|
|
# Firmware host-test gate: run the gcc host unit tests (nvs schema migration,
|
|
# CSI binary-frame serialization, serial_prov JSON parser fuzz) BEFORE the
|
|
# expensive ESP-IDF build so a logic/format-contract regression fails the image
|
|
# build fast. Pure gcc — no IDF toolchain needed; gcc + GNU make ship in this
|
|
# espressif/idf image. `make` propagates the suite's non-zero exit code on any
|
|
# assertion failure, failing the build. This is the gcc harness, NOT idf.py
|
|
# --target linux — see firmware/test/Makefile and the decision record
|
|
# docs/notes/firmware-host-test-approach.md (firmware/main cannot be host-linked).
|
|
RUN make -C test test
|
|
|
|
# Source export.sh to activate IDF toolchain (entrypoint is not called in build stages).
|
|
# set-target must be run explicitly before build even when CONFIG_IDF_TARGET is in sdkconfig.defaults.
|
|
# idf.py build produces build/spaxel-firmware.bin
|
|
SHELL ["/bin/bash", "-c"]
|
|
RUN . $IDF_PATH/export.sh && idf.py set-target esp32s3 && idf.py build && \
|
|
python -m esptool --chip esp32s3 merge_bin \
|
|
--flash_mode dio --flash_freq 80m --flash_size 4MB \
|
|
--output build/spaxel-firmware-merged.bin \
|
|
0x0 build/bootloader/bootloader.bin \
|
|
0x8000 build/partition_table/partition-table.bin \
|
|
0x10000 build/spaxel-firmware.bin
|
|
|
|
# Stage 2: Build the Go binary (cross-platform)
|
|
FROM golang:1.25-bookworm AS builder
|
|
|
|
WORKDIR /app
|
|
|
|
# Copy Go module files first for better caching
|
|
COPY mothership/go.mod mothership/go.sum ./
|
|
RUN go mod download
|
|
|
|
# Copy source code
|
|
COPY mothership/ ./
|
|
|
|
# Copy dashboard files into the mothership cmd/mothership directory for go:embed
|
|
# The go:embed directive in cmd/mothership/main.go references the local dashboard directory
|
|
COPY dashboard/ ./cmd/mothership/dashboard/
|
|
|
|
# Build the binary. CI builds amd64 only (ESP-IDF firmware is x86_64-only),
|
|
# so GOOS/GOARCH are pinned to linux/amd64.
|
|
# CGO_ENABLED=0 because we use pure-Go SQLite (modernc.org/sqlite)
|
|
# -tags=embed enables dashboard embedding via go:embed
|
|
ARG VERSION=dev
|
|
ARG TARGETPLATFORM
|
|
RUN CGO_ENABLED=0 \
|
|
GOOS=linux GOARCH=amd64 \
|
|
go build \
|
|
-ldflags="-s -w -X main.version=${VERSION}" \
|
|
-tags=embed \
|
|
-o spaxel ./cmd/mothership
|
|
|
|
# Also build the CSI simulator so the same image can run synthetic-node load
|
|
# against a deployed mothership (used by the in-cluster simulator workload).
|
|
RUN CGO_ENABLED=0 \
|
|
GOOS=linux GOARCH=amd64 \
|
|
go build \
|
|
-ldflags="-s -w" \
|
|
-o spaxel-sim ./cmd/sim
|
|
|
|
# Stage 3: Minimal runtime image - distroless nonroot
|
|
# Dashboard is embedded in the Go binary via go:embed, not copied as files
|
|
FROM gcr.io/distroless/static-debian12:nonroot
|
|
ARG TARGETARCH=amd64
|
|
|
|
# Copy the binary (dashboard is embedded via go:embed)
|
|
COPY --from=builder /app/spaxel /spaxel
|
|
|
|
# CSI simulator binary — invoked via an explicit command override in the
|
|
# simulator workload; the default ENTRYPOINT still runs the mothership.
|
|
COPY --from=builder /app/spaxel-sim /spaxel-sim
|
|
|
|
# Bake ESP32 firmware into the image so the mothership can seed it on first run.
|
|
# The mothership copies /firmware/*.bin → /data/firmware/ at startup if not present.
|
|
# Firmware is only included on amd64 builds (ESP-IDF is x86_64-only).
|
|
# For non-amd64 builds, the placeholder from firmware-builder stage is included.
|
|
COPY --from=firmware-builder /project/build/spaxel-firmware-merged.bin /firmware/spaxel-firmware.bin
|
|
|
|
VOLUME ["/data"]
|
|
|
|
# Expose HTTP/WebSocket port
|
|
EXPOSE 8080
|
|
|
|
# Run as non-root
|
|
ENTRYPOINT ["/spaxel"]
|