| Age | Commit message (Collapse) | Author |
|---|
|
Accept the defined silence value and round ordinary volume requests to the nearest supported resolution step within the cached range. Add tests for silence, clamping, and step alignment.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Validate descriptor lengths before typed access, keep supported alternate settings when neighboring formats are unsupported, and reject MIDI 1.0 and 2.0 streaming interfaces from the Audio host parser. Extend parser coverage for malformed and mixed descriptor layouts.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Enforce a shared capture/playback sample rate only while the opposite stream is running. This lets both stopped streams be configured in either order while preventing a live shared clock from changing underneath a transfer.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Move control-transfer bookkeeping out of USB memory and overlay mount-only Clock RANGE storage with cache-aligned runtime control payload buffers. Keep explicit feedback separate because it may overlap runtime requests.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Keep the fractional packet accumulator when a new explicit-feedback value is latched. Repeated feedback updates no longer bias high-speed fractional rates toward alternating undersized and oversized packets.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Add protocol-selectable UAC1/UAC2 parsing, UAC2 terminal and Feature Unit topology, Clock Source discovery, sampling-frequency ranges, and protocol-specific stream controls. Cover UAC2 playback, capture, control discovery, and malformed descriptors with unit tests.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Retain every valid discrete rate in descriptor order, including duplicate public tuples. Alternate settings remain independently selectable without disruptive validation or deduplication during enumeration.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Store each Audio Streaming alternate setting once and flatten its sample rates only at the public API boundary. Keep feedback and pacing state playback-only, eliminating duplicate configuration fields from capture streams.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Signed-off-by: HiFiPHile <[email protected]>
|
|
Signed-off-by: HiFiPHile <[email protected]>
|
|
|
|
Inspect each associated Feature Unit during mount, cache master mute support and the common MIN, MAX, and RES volume range, and ignore units with neither control. Add typed asynchronous and synchronous mute and volume APIs and demonstrate them in the host example.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Check descriptor bounds, minimum lengths, frequency counts, and interface types with the host validation helpers before accessing fields. Malformed UAC1 functions now fail enumeration without out-of-bounds reads.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Keep the sampling-frequency control flag in the alternate-setting parser state shared by its single data endpoint. This supports either descriptor order without separate pending and unassigned endpoint flags.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Treat isochronous bInterval as a power-of-two exponent for both full and high speed, using the appropriate frame unit. Packet sizing and fractional scheduling now use the actual service interval.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Retain an audio function when at least one direction has a supported configuration, but release the tentative instance when neither direction does. Unsupported and MIDI-only functions no longer consume host slots or emit mount callbacks.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Keep one unresolved Feature Unit candidate per stream direction while parsing AudioControl entities. Duplex functions are now associated correctly even when both Feature Units precede their USB terminals.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Require the UAC1 AudioControl protocol before allocating an instance. The current parser assumes UAC1 descriptor layouts, so accepting UAC2 would misinterpret entities and consume a host slot.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Route asynchronous activation and sampling-frequency failures through the stream error callback. Keep a running stream active when SET_INTERFACE alt 0 cannot be submitted so stop can be retried without diverging from device state.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Collect one audio data endpoint for each alternate setting and ignore explicit feedback endpoints until feedback scheduling is implemented. Validate the maximum packet for every discrete rate against the endpoint, transfer buffer, and FIFO.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Parse the contiguous AudioStreaming interfaces by class and subclass instead of copying baInterfaceNr. This removes the UAC1-only collection limit and prevents MIDI AudioControl interfaces from leaving a half-allocated Audio Host instance.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Make configuration a synchronous local operation that selects and opens the endpoint. Start now activates the alternate setting and sets the sampling frequency afterward, including every restart, so devices cannot reset the selected rate with SET_INTERFACE.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Keep capture and playback transfers continuously armed from their completion callbacks. Capture overwrites the oldest complete frames when full, while playback sends silence on underrun without consuming partial queued audio.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Bring the audio work onto the current host core and build files before applying the remaining review fixes.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Record terminal and Feature Unit links while parsing the AudioControl block, then resolve the stream association after all entities are known. Common capture and playback descriptor orders are both supported.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Collect only format descriptors with discrete sampling frequencies. The configuration API exposes concrete tuples and cannot safely represent a continuous range as one selectable rate.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Convert full-speed bInterval values to the scheduler microframe timebase and size transfers for the complete service interval. Endpoints with bInterval greater than one no longer run at a one-frame cadence.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Associate one directly connected Feature Unit with each capture and playback stream. This lets applications control microphone and speaker paths independently.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Reset Feature Unit state when parsing a new audio function. Devices without a matching unit must not inherit an ID collected from an earlier interface or device.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Keep Feature Unit request data in driver-owned storage until asynchronous completion and use the width defined for each control. This avoids overlapping request state and malformed one-byte controls.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Clear the started state only after SET_INTERFACE alt 0 succeeds. A failed control transfer can then be retried instead of leaving software and device state inconsistent.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Close the previously selected HCD endpoint before opening another alternate setting. This prevents stale endpoint size and controller state from surviving a format change.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Apply the host memory section and alignment attributes to audio transfer storage. This keeps buffers accessible to controllers that require DMA-capable memory.
Signed-off-by: HiFiPHile <[email protected]>
|
|
Associate UAC1 AudioStreaming endpoints by their usage type. Keep data and implicit-feedback IN endpoints as streams while leaving explicit feedback endpoints unopened until feedback scheduling is supported.
Signed-off-by: HiFiPHile <[email protected]>
|
|
vendor_host.c/.h implemented a CFG_TUH_VENDOR class driver that no example, board
or test ever enabled: usbh's driver table entry was compiled out everywhere, and
the six tusb_config.h files that mentioned the macro all set it to 0. Maintainer
call - dead code, not a shrinking of supported classes.
Removes the sources, the usbh driver-table entry, the CFG_TUH_VENDOR default in
tusb_option.h, the tusb.h include, both build-system source lists, the rp2040
family.cmake entry and the IAR project template rows.
|
|
|
|
This adds more standard terminal types that can
be used in descriptors to improve end user experience
when operating system can present more acurate
image for audio device terminal.
Signed-off-by: Jerzy Kasenberg <[email protected]>
|
|
Fix HFP HIL reliability issue
|
|
midi2 device: complete the UMP stream discovery responder
|
|
clamp committed video payload size to streaming ep buffer
|
|
Co-authored-by: Copilot Autofix powered by AI <[email protected]>
Signed-off-by: HiFiPhile <[email protected]>
|
|
Signed-off-by: HiFiPhile <[email protected]>
|
|
Text replies resume instead of dropping their tail packets, which used
to leave a Start/Continue sequence without an End. A new Function Block
Discovery now merges with a pending one instead of replacing it.
|
|
Adds the Device Identity Notification with an app callback, MIDI-CI
version and SysEx8 stream count in FB Info, honors the Endpoint
Discovery filter bitmap, and paces discovery replies by TX FIFO room.
|
|
A false return from tud_usbtmc_start_bus_read() here does not mean arming
failed: it means the endpoint is already armed, either because the
application re-armed it from its trigger callback or because a transfer is
still queued (usbd_edpt_xfer() reports failure when the endpoint is busy).
Both cases end in STATE_IDLE, so the state cannot disambiguate them either,
and stalling on the result would halt a healthy endpoint.
Co-Authored-By: Claude Opus 5 <[email protected]>
|
|
A single USB488 TRIGGER message left the bulk-OUT endpoint un-armed, so the
host's next bulk-OUT transfer timed out. The trigger itself succeeded
silently, so the failure surfaced on a later, unrelated command; only a
USBTMC device clear recovered it. The bundled examples/device/usbtmc
reproduced this as shipped.
Every other branch of the STATE_IDLE dispatch in usbtmcd_xfer_cb() leaves
the endpoint in a defined state: it either transitions out of STATE_IDLE so
a later tud_usbtmc_start_bus_read() can re-arm it, or it stalls and lets the
CLEAR_FEATURE(ENDPOINT_HALT) handler recover it. USBTMC_MSGID_USB488_TRIGGER
did neither, and because the state stayed STATE_IDLE, even an application
following the contract documented in usbtmc_device.h got a silent no-op from
tud_usbtmc_start_bus_read().
Transition to STATE_NAK so the re-arm can take effect, and stall the
endpoint when trigger is unsupported or the application callback rejects it,
matching the existing handling for messages the driver cannot process. The
callback result is deliberately not wrapped in TU_VERIFY(), which would
return before the stall/re-arm and reintroduce the same hang.
Since the driver now re-arms after a trigger, drop tud_usbtmc_msg_trigger_cb
from the list of callbacks after which the application must do so.
Fixes #3821
Co-Authored-By: Claude Opus 5 <[email protected]>
|
|
Provide a high-level audio streaming API over UAC 1.0 devices while
keeping the USB topology private: applications select supported
{format, sample_rate, channels} configurations per logical stream, and
the driver owns the mapping to AS interface, alternate setting, and
endpoint.
- One logical stream per direction per instance; multiple AS interfaces
and alternate settings in a direction are merged into the stream's
configuration list (discrete tuples; continuous ranges exposed as a
single configuration at the top rate)
- Asynchronous tuh_audio_configure(): SET_INTERFACE to the selected
alternate setting, open/reconfigure the endpoint, set the sampling
frequency when supported, initialize the FIFO and packet scheduler,
then invoke the completion callback
- Frame-based FIFO streaming: tuh_audio_read()/tuh_audio_write() queue
whole frames; the driver owns transfer replenishment and fractional
packet scheduling (44.1 kHz pays back the 0.1 frame/ms remainder via
an accumulator for exact average pacing)
- tuh_audio_start()/tuh_audio_stop() activate/deactivate the stream
interface through SET_INTERFACE (alt n / alt 0)
Driver correctness fixes:
- Parse only the AC header's interface collection; MIDI Streaming and
other subclasses are skipped
- Keep every discrete format as a separate configuration; endpoints are
opened only for the alternate setting selected by tuh_audio_configure()
- Check tuh_interface_set() return values and SET_INTERFACE transfer
results instead of ignoring failures
- Validate instance state, direction, buffers, and frame counts in every
transfer API
- Feature Unit requests use the control's real width (mute/AGC/loudness
1 byte, others 2 bytes) and convert multibyte values to host order
- Failed/stalled/aborted isochronous transfers reach only the error
callback, never the capture/playback callbacks
The audio_host example uses the new API: 48 kHz stereo by default,
automatic stream restart on error callbacks, a sine test tone on the
playback stream, and periodic mic-only / spk-only / echo phase switching.
|
|
morse-cedricvandenbergh/fix/ncm-link-state-notify-retry
ncm: retry link-state notification, fix carrier lost on collision
|
|
tud_network_link_state() delivered the NETWORK_CONNECTION notification
edge-triggered and fire-once: if a previous notification was still in
flight, notification_xmit() returned early and the notification for the
new link state was never queued. Because link_is_up is committed before
the send, the host could be left reporting a stale carrier state - e.g.
a permanent NO-CARRIER after a link up. The notification state was also
mutated from both the caller and the notify xfer-completion callback
with no serialisation, so on RTOS ports where tud_network_link_state()
runs in a task other than tud_task() the two could race.
Defer the whole link-state update onto the usbd task, so it can no
longer race the completion callback. A collision with an in-flight
notification is resolved by re-arming notification_xmit_state and
letting the existing completion callback drive it forward on the next
xfer completion, rather than adding a separate pending/retry flag.
A link toggle does not change the link speed, so strictly only the
NETWORK_CONNECTION notification needs (re)sending, but reusing the
existing speed-then-connection state machine keeps the fix on a single,
already-serialised code path.
Closes #3760
|
|
guard
The comment above audiod_tx_packet_size() states flow control needs a FIFO
of at least 4*Navg, but the guard tests nominal_size[1] <= fifo_depth * 4 -
true for any FIFO larger than a quarter packet - instead of
nominal_size[1] * 4 <= fifo_depth. As written, flow control engages on
FIFOs far below its own documented minimum, where the depth/2 setpoint sits
within one packet of empty and the packet_size = 0 branch (a zero-length
packet, i.e. an audible 1 ms dropout for audio-class hosts) is reachable
from ordinary scheduling jitter rather than only from gross clock
deviation. With the guard corrected, undersized FIFOs fall back to the
plain min(count, max) path as intended.
|
