tinyusb.git/examples, branch claude/adoring-pasteur-kbaFa

bare_api: cast config-descriptor buffer via uintptr_t instead of suppressing V641

2026-06-12T09:15:43Z

temp_buf is a uint16_t[128] reused to hold the GET_DESCRIPTOR(Configuration) wire-format blob; casting it to tusb_desc_configuration_t* to read the 9-byte header tripped PVS V641 (buffer size not a multiple of the element size). Route the cast through uintptr_t so the (correct) header read no longer trips the size-ratio heuristic, dropping the inline //-V641 suppression. Verified: pico examples rebuild; single-TU PVS re-run on bare_api/main.c is clean (no V641, no MISRA pointer/integer-cast finding).

clean up PVS-Studio static-analysis findings for raspberry_pi_pico

2026-06-11T23:34:30Z

Resolves all TinyUSB-owned alerts reported by the CI PVS-Studio job (static_analysis.yml, run with --security-related-issues) on the raspberry_pi_pico board: 0 remaining in src/ and examples/. Genuine fixes: - ncm_device: validate wNdpIndex against sizeof(nth16_t), not the pointer size sizeof(nth16) (4 bytes) — the latter under-checks the NTB header (V568). - tusb: drop the redundant `ff_buf != NULL && ff_bufsize > 0` guard in tu_edpt_stream_init(); the early return already guarantees it (V560). - midi_host: bounds-check idx in tuh_midi_itf_get_info() instead of the always-true `&_midi_host[idx]` pointer (V560). - examples: fully initialize resolutions_per_format / frame_num / interval_ms arrays instead of leaving trailing elements implicitly zero (V1009). False positives suppressed at the cause: - usbd/usbh: hide the weak dcd_deinit()/hcd_deinit() stubs from the analyzer with #ifndef PVS_STUDIO. PVS analyzes one TU at a time and binds the call to the always-false weak stub (it cannot model the linker selecting the port's strong definition), then reports the cleanup loop after TU_ASSERT(...deinit()) as unreachable (V779). False positives suppressed locally (inline //-V or .pvsconfig): - .pvsconfig: V501 (HID descriptor macros), V763 (rhport override), V785 (audio function-index switch), V1044 (hardware poll loops). - inline //-V for config-dependent or intentional constructs: V512, V514-style contiguous clears, V547, V557, V560, V614, V619, V641, V1008, V1037, V1048, V1086. Verified: all examples build for raspberry_pi_pico; ceedling test:all passes (60/60); re-run of the CI-exact PVS invocation reports zero TinyUSB-owned findings.

Merge remote-tracking branch 'origin/master' into stm32c5

2026-06-11T10:20:15Z

# Conflicts: # README.rst

HIL: replace build.flags_on with named build variants (#3687)

2026-06-11T01:16:43Z

* test/hil: replace build.flags_on with named variant schema Boards declare build variants as `variant: [{name, flags}]` instead of `build.flags_on`. The variant `name` is the build dir (cmake-build-) and the HIL report row; `flags` is the raw CFLAGS string (-D...=1) injected via CFLAGS_CLI. No `variant` => a single build named after the board. - build.py: --build-name (dir) + --cflag= (raw CFLAGS, repeatable, =form survives the matrix's shell word-splitting); drop -f1/CFLAGS wrapping. - hil_ci_set_matrix.py: emit one build arg per variant. - hil_test.py: iterate variants; report row + build dir = variant name. - hil_ci.sh: copy all cmake-build-* dirs for -b runs. - get_deps.py: accept (ignore) --build-name/--cflag from matrix args. - tinyusb.json: migrate all 6 flags_on boards to variant. * board_test: park CI build with busy spin instead of wfe

Merge pull request #3690 from hathach/claude/board-test-idle-park

2026-06-10T11:04:54Z

hil: park boards with idle board_test instead of erasing flash

Merge pull request #3657 from hathach/usbh-add-control-queue

2026-06-01T05:42:03Z

Add control transfer fifo for host stack

ultrareview nits: keep xfer_result table in sync, hoist blinky loop

2026-06-01T03:05:17Z

- src/tusb.c: extend tu_str_xfer_result[] with "ABORTED" and "INVALID" to match the new enum size. Not reachable today (no HCD posts those values through hcd_event_xfer_complete), but keeps the enum/table invariant intact so future HCDs that surface ABORTED don't index OOB. - examples/dual/dynamic_switch/src/main.c: apply the same while(1) hoist already done for cdc_task / print_devinfo_task to led_blinking_task. On OS_NONE the loop returned mid-iteration, which on first call could fire multiple back-to-back toggles while start_ms (initially 0) caught up to uptime. Co-Authored-By: Claude Opus 4.7

esp32 build fixes

2026-05-30T09:46:18Z

Signed-off-by: HiFiPhile

dynamic_switch: hoist while(1) out for OS_NONE

2026-05-29T17:52:22Z

Sonar flagged the loop body as executing only once on OS_NONE because the OS_NONE branch returns inside the first iteration (main() drives the task again). Make the while(1) conditional on RTOS so the OS_NONE build is a straight-line function with no misleading loop. Co-Authored-By: Claude Opus 4.7

midi2_host: silence IAR Pe550 for midi2_idx

2026-05-29T17:46:16Z

The variable is set in mount/umount callbacks but not read elsewhere in the example (rx_cb already receives idx as a parameter). IAR treats Pe550 as an error under --warnings_are_errors. Tag it TU_ATTR_UNUSED so the example still shows the pattern of tracking the device index without erroring on unused-set. Co-Authored-By: Claude Opus 4.7