feat(usb_host): HID Host add suspend/resume events

[peter-marcisovsky]

Description

This MR adds Suspend/Resume events for HID class driver as a follow-up for the Global/Suspend resume MR in esp-idf

Changes

Added Suspend and Resume HID Host interface events

// For backward compatibility with IDF versions which do not have suspend/resume api
#ifdef USB_HOST_LIB_EVENT_FLAGS_AUTO_SUSPEND
#define HID_HOST_SUSPEND_RESUME_API_SUPPORTED
#endif

As the global Suspend issued by the USB Host Halts and flushes all EPs of all connected devices. The HID Host driver must submit the in transfer upon Resuming the root port (similar to hid_host_device_start() when a new device is connected)

Related

• Closes Suspend/Resume support for HID Class driver

Testing

• Added target tests to test HID Host reaction to suspend/resume events

---

Checklist

Before submitting a Pull Request, please ensure the following:

• 🚨 This PR does not introduce breaking changes.
• All CI checks (GH Actions) pass.
• Documentation is updated as needed.
• Tests are updated or added as necessary.
• Code is well-commented, especially in complex areas.
• Git history is clean — commits are squashed to the minimum necessary.

---

Note

Add global suspend/resume handling and events to HID Host, with state management updates and comprehensive tests.

• HID Host core (driver):
  • Add HID_INTERFACE_STATE_SUSPENDED and last_state to track suspend/resume.
  • Implement suspend/resume helpers: hid_host_suspend_interface(), hid_host_resume_interface(), plus device-level iterators to broadcast events per interface.
  • Extend client event callback to handle USB_HOST_CLIENT_EVENT_DEV_SUSPENDED/DEV_RESUMED and log others via switch.
  • Update disable/close paths to work when interface is suspended (EP clear logic, state transitions).
• API/Header (include/usb/hid_host.h):
  • Add conditional HID_HOST_SUSPEND_RESUME_API_SUPPORTED (gated by USB_HOST_LIB_EVENT_FLAGS_AUTO_SUSPEND).
  • Add new interface events: HID_HOST_INTERFACE_EVENT_SUSPENDED and HID_HOST_INTERFACE_EVENT_RESUMED.
• Tests (test_app):
  • Refactor test event queue to carry both driver and interface events; initialize/cleanup queue defensively.
  • Add PM tests: suspend_resume_basic, auto_suspend_timer, resume_by_transfer_submit, sudden_disconnect_suspended_device.
  • Add error-handling tests for opening/starting devices while suspended; minor logging improvements.
  • USB lib task handles USB_HOST_LIB_EVENT_FLAGS_AUTO_SUSPEND by suspending root port.
• Docs:
  • Update CHANGELOG.md under [unreleased] with global suspend/resume support.

^{Written by Cursor Bugbot for commit 2aad240. This will update automatically on new commits. Configure here.}

[cursor]

This is the final PR Bugbot will review for you during this billing cycle

Your free Bugbot reviews will reset on November 20

Details

Your team is on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle for each member of your team.

To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.

[cursor]

Bug: Transfer resubmitted for interfaces not in ACTIVE state

The hid_host_resume_interface function unconditionally resubmits the transfer if in_xfer is not NULL, regardless of the interface's previous state. This causes transfers to be submitted for interfaces that were in HID_INTERFACE_STATE_READY state before suspension, violating the state machine where transfers should only be active when the interface is in HID_INTERFACE_STATE_ACTIVE state. The transfer should only be resubmitted when last_state equals HID_INTERFACE_STATE_ACTIVE.

 

[cursor]

Bug: Unchecked return values in suspend/resume device functions

The hid_host_suspend_interface return value is ignored in hid_host_device_suspended, allowing the suspended event callback to be delivered even if the suspend operation failed. This creates inconsistency between the actual interface state and the events delivered to the user. The same issue exists in hid_host_device_resumed at lines 691-692 where hid_host_resume_interface errors are ignored.

 

[roma-jam]

One more question from my side for now: