Conversation
There was a problem hiding this comment.
Pull request overview
This PR adds a new device-side “usbtest” example (to act as a peer for the Linux kernel usbtest/testusb battery), integrates a host-side runner into HIL, and includes a broad set of DCD/core fixes uncovered by running the battery across many device ports.
Changes:
- Add
examples/device/usbtestplus vendor-class extensions (altsettings + optional bulk/int/iso endpoints) to support the Linuxusbtestsource/sink protocol. - Add
test/hil/usbtest.pyrunner and wire it into the HIL harness/reporting and rig configuration. - Fix multiple device-controller drivers and USBD core behaviors exposed by the usbtest battery (stall/clear-stall, toggle reset, iso support, FIFO readiness, etc.).
Reviewed changes
Copilot reviewed 35 out of 37 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| test/hil/usbtest.py | New host-side runner for binding usbtest + executing/parsing the testusb battery. |
| test/hil/tinyusb.json | Register CH32V307 variants and adjust rig board config/skip list. |
| test/hil/tinyusb-sudoer | Add sudoers allowlist template for usbtest binding/recovery workflows. |
| test/hil/hil_test.py | Integrate device/usbtest into HIL, add metric-aware failures and reporting tally fixes. |
| src/portable/wch/dcd_ch32_usbhs.c | Fix clear-stall behavior to re-queue armed OUT transfers (toggle reset case). |
| src/portable/wch/dcd_ch32_usbfs.c | Add/enable iso support details, per-direction iso state, EP3 sizing, and toggle handling fixes. |
| src/portable/wch/ch32_usbfs_reg.h | Add EP3 1023B capability macro for relevant WCH MCUs. |
| src/portable/st/stm32_fsdev/dcd_stm32_fsdev.c | Refine clear-stall so it doesn’t disarm VALID endpoints unintentionally. |
| src/portable/renesas/rusb2/dcd_rusb2.c | Bound FIFO-ready spins, add ZLP deferral, improve control OUT flow control, and fix clear-stall re-arm logic. |
| src/portable/raspberrypi/rp2040/rp2040_usb.c | Prevent iso endpoints from being double-buffered into neighboring DPRAM. |
| src/portable/raspberrypi/rp2040/dcd_rp2040.c | Re-issue active transfers on clear-stall toggle reset so claims complete properly. |
| src/portable/nxp/lpc_ip3511/dcd_lpc_ip3511.c | Fix stall semantics and implement iso alloc/activate handling. |
| src/portable/nordic/nrf5x/dcd_nrf5x.c | Add errata 199 workaround latch + implement iso alloc/activate path. |
| src/portable/microchip/samd/dcd_samd.c | Implement iso alloc/activate to reserve bank sizes and enable ISO endpoints. |
| src/portable/mentor/musb/dcd_musb.c | Prevent IN ISR from loading disarmed transfers; flush FIFO before stalling. |
| src/device/usbd.c | Harden GET_STATUS handling and endpoint feature validation; adjust vendor control callback wiring; refine clear-stall claim release. |
| src/common/tusb_mcu.h | Update endpoint-close vs iso-alloc capability documentation/deprecation notes. |
| src/class/vendor/vendor_device.h | Add config gates and APIs for optional interrupt/iso endpoints + altsetting support. |
| src/class/vendor/vendor_device.c | Implement vendor altsettings logic and add non-buffered helpers for bulk/int/iso endpoints. |
| src/class/audio/audio_device.c | Remove prior “clear busy” workaround comment in favor of standardized endpoint state clearing. |
| lib/rt-thread/port/msc_device_port.c | Bound INQUIRY string copies to avoid overflow. |
| hw/bsp/ch32v30x/family.c | Implement board_get_unique_id() for CH32V30x using ESIG UID. |
| hw/bsp/ch32v20x/family.c | Replace interrupt handlers with nested-safe naked ISRs for QingKe HWSTK scenario. |
| examples/device/video_capture/src/main.c | Fix VIDEO_STACK_SIZE typo. |
| examples/device/video_capture_2ch/src/main.c | Fix VIDEO_STACK_SIZE typo. |
| examples/device/usbtest/src/usb_descriptors.h | New usbtest descriptor/tier constants and per-MCU endpoint sizing rules. |
| examples/device/usbtest/src/usb_descriptors.c | New FS/HS descriptors (alt0 empty, alt1 endpoints) and string descriptors for usbtest. |
| examples/device/usbtest/src/tusb_config.h | New example TinyUSB config enabling vendor bulk/int/iso + altsettings and manual RX arming. |
| examples/device/usbtest/src/main.c | New device firmware implementing source/sink pumps + ctrl_out protocol. |
| examples/device/usbtest/src/CMakeLists.txt | ESP-IDF component registration for usbtest example. |
| examples/device/usbtest/skip.txt | Skip list for unsupported/unsuitable MCUs. |
| examples/device/usbtest/README.md | Documentation for tiers, protocol, and running via test/hil/usbtest.py. |
| examples/device/usbtest/Makefile | Make build integration for the new example. |
| examples/device/usbtest/CMakePresets.json | Presets include for board presets (CMake). |
| examples/device/usbtest/CMakeLists.txt | CMake build integration for the new example. |
| examples/device/CMakeLists.txt | Add usbtest to the device example list. |
| .claude/skills/usb-recover/SKILL.md | New rig USB recovery documentation. |
| .claude/skills/usb-recover/scripts/usb_recover.sh | New restricted recovery helper script (sudoers-friendly). |
| .claude/skills/usb-debug/SKILL.md | New rig USB dynamic-debug documentation. |
| .claude/skills/usb-debug/scripts/usb_dyndbg.sh | New restricted dynamic-debug toggle helper script (sudoers-friendly). |
| .claude/skills/hil/SKILL.md | Add guidance to stop/start the Actions runner on ci during manual hardware use. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Size Difference ReportBecause TinyUSB code size varies by port and configuration, the metrics below represent the averaged totals across all example builds. Note: If there is no change, only one value is shown. Changes >1% in size
Changes <1% in size
No changes
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 099b6538dc
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Code reviewNo issues found. Checked for bugs and CLAUDE.md compliance. |
1 similar comment
Code reviewNo issues found. Checked for bugs and CLAUDE.md compliance. |
|
| target | .text | .rodata | .data | .bss | total | % diff |
|---|---|---|---|---|---|---|
| raspberrypi_zero/webusb_serial | 48,920 → 49,384 (+464) | 1,087 → 4,719 (+3,632) | — | — | 50,007 → 54,103 (+4,096) | +8.2% |
| raspberrypi_zero/audio_4_channel_mic | 57,216 → 57,592 (+376) | 1,591 → 5,311 (+3,720) | — | — | 58,807 → 62,903 (+4,096) | +7.0% |
| ch32v103c_bluepill/dfu_runtime | 8,556 → 8,668 (+112) | — | — | 2,880 → 1,800 (-1,080) | 13,900 → 12,932 (-968) | -7.0% |
| raspberrypi_zero/dynamic_configuration | 61,420 → 61,816 (+396) | 939 → 4,639 (+3,700) | — | — | 62,359 → 66,455 (+4,096) | +6.6% |
| raspberrypi_cm4/cdc_uac2 | 61,248 → 61,568 (+320) | 1,320 → 5,096 (+3,776) | — | — | 62,568 → 66,664 (+4,096) | +6.5% |
| ch32v103c_bluepill/hid_generic_inout | 9,528 → 9,648 (+120) | — | — | 3,084 → 2,004 (-1,080) | 15,080 → 14,120 (-960) | -6.4% |
| ch32v103c_bluepill/hid_multiple_interface | 10,368 → 10,484 (+116) | — | — | 2,960 → 1,880 (-1,080) | 15,792 → 14,828 (-964) | -6.1% |
| ch32v103c_bluepill/hid_boot_interface | 10,388 → 10,504 (+116) | — | — | 2,960 → 1,880 (-1,080) | 15,812 → 14,848 (-964) | -6.1% |
| ch32v103c_bluepill/hid_composite | 10,612 → 10,732 (+120) | — | — | 2,948 → 1,868 (-1,080) | 16,028 → 15,068 (-960) | -6.0% |
| portenta_c33/uac2_headset | 15,420 → 16,340 (+920) | — | — | — | 15,468 → 16,388 (+920) | +5.9% |
Hardware-in-the-loop (HIL) Test Reporthfp-iar✅ 42 passed · ❌ 14 failed · ⚪ 0 skipped · blank not run
hfp.json✅ 39 passed · ❌ 14 failed · ⚪ 0 skipped · blank not run
tinyusb-esp.json✅ 21 passed · ❌ 3 failed · ⚪ 0 skipped · blank not run
tinyusb.json✅ 314 passed · ❌ 25 failed · ⚪ 11 skipped · blank not run
|
There was a problem hiding this comment.
💡 Codex Review
tinyusb/src/class/vendor/vendor_device.c
Lines 152 to 155 in 6d707a1
Include new vendor endpoints in mounted checks
With the new interrupt/iso endpoint support, a valid vendor interface no longer necessarily has bulk endpoints. For an interrupt-only descriptor, or an altsetting that exposes only interrupt/iso endpoints, vendord_open() records the new endpoint fields but tud_vendor_n_mounted() still checks only the bulk streams/endpoints, so typical application code guarded by tud_vendor_mounted() will never arm or write those endpoints even though the interface is configured.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
tinyusb/src/class/vendor/vendor_device.c
Line 155 in b7acb73
Include optional vendor endpoints in mounted checks
When CFG_TUD_VENDOR_TXRX_BUFFERED=0 and a descriptor uses the new interrupt-only vendor support (the open path explicitly allows bulk to be absent), vendord_open() records ep_int_in/out but leaves ep_in/out at zero, so tud_vendor_mounted() reports false even though the interface is configured and its interrupt endpoints are open. Applications that gate tud_vendor_int_read_xfer() or tud_vendor_int_write() on the mounted predicate will never arm/service these endpoints; include the optional interrupt/iso endpoint fields in this predicate.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 24a92e4831
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
# Conflicts: # .claude/skills/hil/SKILL.md # test/hil/hil_test.py
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a686c70631
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| static inline void vendord_abort_ep(uint8_t rhport, uint8_t ep_addr) { | ||
| if (ep_addr) { | ||
| usbd_edpt_stall(rhport, ep_addr); | ||
| usbd_edpt_clear_stall(rhport, ep_addr); |
There was a problem hiding this comment.
Fresh evidence beyond the earlier altsetting-abort thread is that several updated DCD clear-stall paths now intentionally preserve an armed OUT transfer for toggle-reset recovery (for example dcd_rusb2.c:989 checks _dcd.pipe[num].queued, and WCH checks xfer->valid). This helper's stall+clear sequence therefore is not an abort on those ports: switching a vendor interface from an endpoint altsetting to alt 0 while a read is armed can leave the hardware endpoint ACKing after the class zeros ep_out, so a zero-bandwidth altsetting can still consume packets and any late completion no longer matches the selected endpoints. Use a real DCD abort/close path or clear the DCD transfer state before the clear-stall step.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
Acknowledged -- you're right that vendord_abort_ep's stall+clear-stall sequence isn't a true abort on ports that intentionally keep an armed OUT transfer queued through clear-stall for toggle-reset recovery (e.g. dcd_rusb2.c:989's _dcd.pipe[num].queued check, WCH's xfer->valid). This is a known limitation and is being deferred to a separate follow-up PR (a real DCD-level abort/close path) rather than fixed here.
- dcd_edpt_stall flushes the primed buffer (ENDPTFLUSH), but the aborted transfer's dQH overlay can be left ACTIVE with mid-transfer state; the next prime after clear-halt then resumes the stale overlay instead of loading the fresh qtd, so post-halt IN reads return mid-buffer data (usbtest case 13 'buf[32] = 56 (not 0)', with case 18 failing downstream of the same corruption in the full battery). qhd_start_xfer now clears overlay.active alongside overlay.halted before linking the new qtd. - test/hil(hfp): drop lpcxpresso43s67's device/usbtest skip - the historical first-case wedge no longer reproduces on this branch, and with the overlay fix the board runs 30/30 on its Fresco xHCI host (previously 28/30 with deterministic case 13/18 failures). mimxrt1064_evk (imxrt dcache path) 30/30 regression-clean. - docs(hil skill): document the external hifiphile rig - pool test/hil/hfp.json, SSH-reachable from htpc/ci with no outbound SSH, exercised by the CI hil-tinyusb (hfp.json) job; never run HIL against it during development unless the user explicitly asks. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017TQZrFfU3K4Y198aLsUpBC
…3/p4 and mimxrt1015
edpt_schedule_packets() advanced xfer->buffer past each armed EP0
chunk, so the OUT-complete handler invalidated the cache at the
ADVANCED pointer: one line past the received data. The CPU then read
stale cached bytes instead of the DMA'd packet, and the misplaced
invalidate discarded a dirty line of whatever variable follows the
buffer - random neighbor corruption on every control-OUT data stage.
Found by usbtest ctrl_out (cases 14/21) on espressif_p4_function_ev
with DMA enabled, the first DWC2 target combining buffer DMA with a
data cache: usbd control state wedged after the first control write
(every later request stalled), and one build layout panicked in the
usbd memcpy with a wild pointer.
Rework the EP0 chunk bookkeeping so xfer->buffer always points at the
un-consumed position: the arm no longer advances it; instead the EP0
re-arm paths advance past each completed (full) chunk, invalidating it
first on the OUT side. The final OUT completion invalidates exactly
the received bytes of its last chunk, taken from DOEPDMA ("incremented
on every AHB transaction", databook 7.1.83 - the same semantics the
SETUP path relies on) before dma_setup_prepare() re-targets it. EP0
chunking state (ep0_pending) is now also dropped on bus reset and on
a new SETUP, so a stale latched completion can no longer re-arm EP0
DMA from dead state. No behavior change for targets without dcache.
While root-causing, the FIFO layout was cross-checked against the
DWC2 databook/programming guide v4.20a: the existing GDFIFOCFG
programming (EPInfoBaseAddr = otg_dfifo_depth - 2*ep_count, one SPRAM
word per endpoint direction for buffer DMA) is conformant and needs
no change; the P4 HS instance's reset GDFIFOCFG (0x03800400) merely
reflects a scatter/gather-sized EP_LOC_CNT of 128 that buffer DMA
does not need.
With the fix in place, enable the usbtest battery on the espressif
fleet: tools/build.py allowlists device/usbtest (a plain IDF component
like board_test/video_capture) and both espressif boards' only-lists
gain device/usbtest. Also re-enable device/usbtest on mimxrt1015_evk:
its skip predated the dcd_ci_hs stale-ACTIVE-overlay fix (already on
this branch), which cured the battery that previously killed the
uPD720201 host controller twice (2026-07-11 ROM fw, 2026-07-13 case 27
on fw 2.0.2.6); rig-validated 30/30 three consecutive runs.
Validated on rig (all 30/30): espressif_p4_function_ev(-DMA) (was
22/30 under DMA), espressif_s3_devkitm(-DMA), stm32f723disco(-DMA),
mimxrt1015_evk; p4/s3 slave-mode unaffected (DMA-only code path);
compile-checked stm32h743nucleo +TUD DMA, stm32f407disco,
stm32l476disco (device ports currently on the dead hub).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017TQZrFfU3K4Y198aLsUpBC
hathach
changed the title
… incompatible The MosChip MCS9990 card is physically removed from the rig: delete its cases-11/25 SKIP workaround (and the now-orphaned SKIP accounting) from usbtest.py and refuse to run outright if a DUT ever sits behind one again. usb_recover.sh gains `hub-cycle <busport>`: uhubctl VBUS cycle of the port feeding the device, walking upstream (parent hub -> root port) until it re-enumerates. Verified on the rig: leaf-level recovery (13-4.4 usbtest device) and full walk to the root port on a dead branch. SKILL.md updated for the action and the two-Renesas topology (root-port ppps is real; leaf 1a40:0201 hubs fake their "ganged" switching). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WxUeX4Yn26KibfjvDg2pN9
stm32h533nucleo could never link with IAR: family.cmake points LD_FILE_IAR at linker/stm32h533xx_flash.icf, which did not exist (every sibling H5 variant has one). Surfaced by CircleCI's one-random job picking stm32h533nucleo+IAR (Fatal error[Lc002]). H533 and H523 have identical memory maps (512K flash / 272K RAM; their GCC .ld files differ only in a comment), so the icf is a copy of the H523 one.
Code reviewNo issues found. Checked for bugs and CLAUDE.md compliance. |
8c19913 to
1a89552
Compare
|
@codex review |
| root = Path(f"/sys/bus/usb/devices/usb{int(dev['node'].split('/')[-2])}") | ||
| drv = (root / '../driver').resolve().name | ||
| pci = (root / '..').resolve() | ||
| vid_did = ((pci / 'vendor').read_text().strip(), (pci / 'device').read_text().strip()) |
There was a problem hiding this comment.
Verified on the rig: Path.resolve() has realpath semantics — the usbN symlink is dereferenced before .. is applied, so (root / '..').resolve() lands on the actual parent controller, not back in /sys/bus/usb/devices (that would only happen with lexical normalization, e.g. os.path.normpath). Live check: (Path('/sys/bus/usb/devices/usb11') / '..').resolve() → /sys/devices/pci0000:00/0000:00:1c.0/0000:01:00.0 with vendor/device 0x1912/0x0014 (the Renesas uPD720201) and driver xhci-pci-renesas. The compatibility gate has been correctly identifying controllers and firmware on the HIL rig throughout.
| import re | ||
| import sys | ||
| from pathlib import Path |
There was a problem hiding this comment.
pids: dict[int, list[str]] = {} is a function-local annotated assignment inside main(), not module-level — per PEP 526, local-variable annotations are never evaluated at runtime (CPython ignores the expression entirely), so there is no import-time or call-time subscript of dict and no crash on Python 3.8. Only module/class-level annotations are evaluated; none of those here use PEP-585 generics.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1a89552548
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if (!no_dma && nrf52_errata_199()) { | ||
| NRF_USBD_ERRATA_199_REG = 0x00000082UL; |
There was a problem hiding this comment.
On nRF52840 revisions where nrf52_errata_199() is true, a USB reset or disconnect can arrive while an EasyDMA transfer is active; bus_reset() later clears _dcd without calling edpt_dma_end(), so if no END event accompanies the reset this new NRF_USBD_ERRATA_199_REG = 0x00000082UL latch remains set indefinitely. Since the workaround's purpose is to hold tokens while DMA is pending, leaving it asserted after the transfer was aborted can leave EP0 unable to accept the next SETUP after reset until a DMA completion that may never occur; clear the latch in the reset path as well.
Useful? React with 👍 / 👎.
| { | ||
| "name": "mimxrt1064_evk", | ||
| "uid": "BAE96FB95AFA6DBB8F00005002001200", |
There was a problem hiding this comment.
In this HIL config update, mimxrt1064_evk is moved to boards-skip, but the removed mimxrt1015_evk entry is not re-added anywhere (rg mimxrt1015 test/hil finds no remaining entry). That means hil_ci_set_matrix.py test/hil/tinyusb.json no longer builds or runs this ci_hs board at all, so regressions on the port that this change is meant to cover can land silently; please either restore it under boards or move it to boards-skip with an explicit reason.
Useful? React with 👍 / 👎.
Full-fleet profiling (HIL_PROFILE=1 instrumentation, included) showed each
uPD720201 controller's serialized usbtest battery chain dominates wall time,
and a board whose marginal device port bounces during concurrent batteries
can wedge or kill the controller ("xHCI host not responding to stop endpoint
command"). Every such death traced to mimxrt1015's port (its old "kills the
uPD720201" reputation) - it is removed from the config until recabled;
mimxrt1064's enum-retry stalls were a loose device cable (re-seated).
nrf54lm20dk moves to boards-skip until its failing J-Link probe is replugged.
With the hardware fixed both cards run width-4 batteries plus full flash
churn clean, so scheduling stays simple: two symmetric knobs, flashes and
batteries budgeted per controller.
- schedule_boards(): dispatch boards round-robin across host controllers from
a persisted hint cache (~/.cache/tinyusb-hil/ctrl_cache.json), learned and
merge-on-write refreshed each run (concurrent HIL jobs keep each other's
entries). Only the cached PCI address is consumed - dispatch order and
first-flash budgeting, never battery serialization (batteries resolve live
or fail closed to an all-slot permit).
- HIL_FLASH_PARALLEL (8) and HIL_USBTEST_PARALLEL (4) are budgeted per
controller via lock slots assigned on first sight.
- re-runs: a failed run writes <report dir>/<config>.failed with the exact
re-run spec (--accumulate -b <failed board> -bt <board>:<its failed
tests>) instead of the inverted --skip-board list of everything that
passed; --skip-board is gone, --flasher/--exclude-flasher scope a config
across CI jobs by flasher type (no board names hardcoded in workflows),
and -a/--accumulate merges a re-run into the existing report. The spec is
stamped with GITHUB_RUN_ID and cleared on fresh runs, so a retry can never
consume a spec left behind by a different run's dead or skipped attempt.
- CI: esp-idf firmware builds move out of hil-build into hil-build-esp, and
the esptool-flashed boards run in their own hil-tinyusb-esp job, so the
main hil-tinyusb run starts as soon as the fast toolchains finish instead
of waiting on the slow esp-idf build (an esp toolchain flake previously
skipped the whole rig run). Artifacts are namespaced per toolchain so the
esp job downloads only esp-idf binaries.
- HIL_PROFILE=1: timestamped log lines, per-flash durations, permit-wait
logging, uid->controller map dump for analysis.
- hil_report: per-variant test duration as a dedicated trailing column,
recorded only by full runs.
Validated on the ci rig (fixed seeds 20260716/777, full fleet at 8/4):
738s/780s walls with only known-flake failures and no controller deaths,
vs 1134-1211s serialized-battery baseline.
FS-forced builds on the HS module are not a supported configuration, so the 32-bit access path does not need to exist in full-speed builds: CFG_TUSB_FIFO_HWFIFO_DATA_STRIDE returns to (2 | (TUD_OPT_HIGH_SPEED ? 4 : 0)). Saves 28 bytes of text on FS-only parts (measured on ra4m1_ek cdc_msc); high-speed builds keep both widths for the dual-module (FS+HS ports) case. Build-verified: ra4m1_ek and ra6m5_ek cdc_msc.
vendord_set_alt() aborted bulk/interrupt endpoints of the outgoing altsetting (stall/clear-stall) but only dropped the iso endpoints' tracking: an armed iso transfer stayed live in the dcd with its usbd claim held and no tracked handle to stop it, and its completion fired into an endpoint the class no longer recognizes. Reachable through the usbtest example's alt0 (bulk) <-> alt1 (iso) SET_INTERFACE switching. Track each selected iso endpoint's descriptor (points into the app's static descriptor set) and deactivate on de-selection: with the iso-alloc API re-activation is the abort/scrub primitive (resets ep_status, aborts the stale transfer); without it usbd_edpt_close does, and the next selection re-opens. Iso cannot be stalled like bulk/interrupt, hence the separate path. Build-verified: usbtest for stm32f072disco, ra4m1_ek, raspberry_pi_pico (iso-alloc) and ch32v307v_r1_1v0 (close API).
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a2f0c3c1ba
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| 'device/printer_to_cdc', | ||
| 'device/midi_test', | ||
| 'device/mtp', | ||
| 'device/usbtest', # cafe:4010, unique PID; runs the Linux testusb tier-4 battery via usbtest.py |
There was a problem hiding this comment.
When HIL is run through test/hil/hil_ci.sh, the remote setup step copies only hil_test.py, pymtp.py, and the JSON config (hil_ci.sh:52-55), not the new test/hil/usbtest.py helper or its recovery helper. Since this line adds device/usbtest to every device board's default run list, remote HIL invocations will flash the firmware and then fail with python3 .../usbtest.py: No such file instead of running the battery; copy the helper(s) in hil_ci.sh before enabling it by default.
Useful? React with 👍 / 👎.
Code reviewNo issues found. Checked for bugs and CLAUDE.md compliance. |
ceedling 1.1.0 (released 2026-07-17) fails this project's mock
preprocessing ('Failed to read _build/test/preprocess/.../raw/*.h for
comment stripping'); reproduced locally with an isolated 1.1.0 install
while 1.0.1 passes all 61 tests on the same tree. Unpin once fixed
upstream.
Summary
Adds
examples/device/usbtest— a device-side peer for the Linux kernel'susbtest.ko/testusbbattery (gadget-zero source/sink protocol) — plus a host runner (test/hil/usbtest.py) and HIL integration. Driving the full 30-case battery (bulk / control / interrupt / isochronous, incl. unlink, halt and data-toggle cases) across the rig surfaced and fixed real bugs in 13 device ports, the vendor class and the usbd core. Every device-capable board on the HIL rig passes 30/30.Device example + vendor class
examples/device/usbtest: vendor interface with bulk/int/iso source-sink pairs, iso in altsetting 1, tiered capability viabcdDevice/USBTEST_TIER, control 0x5b/0x5c write/read-back.class/vendor: optional interrupt endpoints, isochronous + alternate-setting support, manual RX arming for halt recovery; de-selected iso endpoints are deactivated on SET_INTERFACE (stale transfers aborted).usbd: harden endpoint/interface standard requests (GET_STATUS(interface), feature ACKs, GET_STATUS direction check); release endpoint claim on clear-stall.DCD fixes (each from a failing battery case, root-caused on hardware)
queuedflagdcd_edpt_iso_alloc/activate(were false-returning stubs)interruptattr corrupts nested QingKe HWSTK returns); real ESIG unique idHost runner + HIL
test/hil/usbtest.py: tiered battery with per-case verdicts, dmesg capture, kernel-hang auto-recovery,--json; refuses known-incompatible host controllers (uPD720201 needs firmware ≥ 2.02.6.9).test/hil/hil_test.py: per-host-controller flash/usbtest budgeting (8/4) with round-robin dispatch from a learned controller cache; failed runs emit an exact re-run spec (--accumulate -b … -bt …) consumed by CI retries; per-variant duration column in the report.hil-build-esp/hil-tinyusb-espjobs (selected by--flasher esptool, no board names in workflows) so the main HIL run no longer waits on the slowest toolchain.Validation
🤖 Generated with Claude Code