feat(canopen): configurable axis object-offset in Ds402Drive for multi-axis drives

[finger563]

Summary

Ds402Drive addressed a single axis at the standard device-profile indices (controlword 0x6040, statusword 0x6041, ...), so it could not drive the second axis of a multi-axis controller.

Basicmicro's MCP266, for example, mirrors every M1 object at +0x800 for M2 (0x6840, 0x6841, 0x687A, ...). This adds Config::object_offset, applied via a new axis_object() helper.

Details

• The offset is applied only to device-profile objects (0x6000-0x6FFF). Communication and identity objects (< 0x6000, e.g. 0x1000 device type, 0x1008 device name, 0x1018 identity) are shared by the whole device and are never offset.
• object_offset defaults to 0, so existing single-axis behavior is unchanged.
• Two Ds402Drive instances (offset 0 and 0x800) can share one CanopenClient to drive both channels of a dual-axis node.

espp::Ds402Drive m1(client, {.object_offset = 0x000});
espp::Ds402Drive m2(client, {.object_offset = 0x800});

Testing

• Existing canopen host test still passes (ALL TESTS PASSED).
• Exercised on real MCP266 hardware driving both channels' profile-position moves through two offset instances over one client.

🤖 Generated with Claude Code

[Copilot]

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Pull request overview

Adds a configurable per-axis object dictionary index offset to Ds402Drive so a single CANopen node can be controlled as multiple DS402 axes (e.g., second axis mirrored at +0x800).

Changes:

• Introduces Config::object_offset and stores it in Ds402Drive.
• Applies the offset to DS402 device-profile SDO accesses via a new axis_object() helper.
• Updates DS402 read/write helpers (controlword/statusword/modes/targets/profile params) to use the offset mapping.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[github-actions]

✅Static analysis result - no issues found! ✅

[finger563]

Addressed the review comment:

• Offset range: axis_object() now restricts the offset to the inclusive CiA 402 device-profile range 0x6000–0x6FFF. Objects at 0x7000+ (e.g. manufacturer-specific ranges) are no longer offset, matching the documented behavior; comment clarified.

[Copilot]

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

[finger563]

Second round addressed:

• Comment overclaim: reworded — the helper no longer asserts that >= 0x7000 objects are "shared / not axis-relative" (which can be false per device); it now just says those indices "are not offset by this helper and returned unchanged."
• uint16_t wraparound: the offset addition is done in uint32_t, and the offset result is only used when it still fits in uint16_t (<= 0xFFFF); otherwise the original index is returned rather than silently wrapping.

[finger563]

Self-reviewed the full diff — no further issues. Every CiA 402 device-profile object accessor is wrapped with axis_object() while the identity/communication objects (0x1000/0x1001/0x1008/0x1018) are correctly left unoffset; the range is inclusive 0x6000-0x6FFF and the widened-add guards against a uint16_t wrap. Existing host test still passes.

[Copilot]

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

[finger563]

Addressed the latest round:

• Magic numbers → named constants: added OBJ_DEVICE_PROFILE_MIN (0x6000), OBJ_DEVICE_PROFILE_MAX (0x6FFF), OBJ_INDEX_MAX (0xFFFF) and MAX_AXIS_OBJECT_OFFSET (= OBJ_INDEX_MAX - OBJ_DEVICE_PROFILE_MAX) in detail::ds402, and the offset logic uses them.

• Overflow no longer silently uses the wrong axis: the offset is now validated once, in the Ds402Drive constructor. An object_offset greater than MAX_AXIS_OBJECT_OFFSET (i.e. one that could push a device-profile index past the 16-bit index space) is rejected with an error log and offsetting is disabled — rather than silently falling back per-call. With a valid offset, axis_object() can never overflow.

• Host-testable + tested: extracted the mapping into a pure constexpr detail::ds402::apply_axis_offset(index, offset) (which Ds402Drive::axis_object() now delegates to), and added test_axis_offset to the canopen host test covering (1) device-profile indices in 0x6000–0x6FFF are offset (incl. the inclusive boundaries), (2) indices outside the range (identity/comm objects, 0x5FFF, 0x7000) are not offset, and (3) an offset that would overflow returns the index unchanged. ALL TESTS PASSED.