Re: [RFC PATCH 01/16] Documentation: driver: add SoundWire BRA description

[Vinod Koul <vkoul@kernel.org>]

Hi Pierre,

On 07-12-23, 16:29, Pierre-Louis Bossart wrote:
> The Bulk Register Access protocol was left as a TODO topic since
> 2018. It's time to document this protocol and the design of its Linux
> support.

Thanks for letting me know about this thread. At least now with B4 we
can grab threads we are not copied.

> 
> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> ---
>  Documentation/driver-api/soundwire/bra.rst    | 478 ++++++++++++++++++

Can we split the cadence parts of this to bra-cadence.rst that way this
file documents the core parts only

> +Concurrency between BRA and regular read/write
> +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> +
> +The existing 'nread/nwrite' API already relies on a notion of start
> +address and number of bytes, so it would be possible to extend this
> +API with a 'hint' requesting BPT/BRA be used.
> +
> +However BRA transfers could be quite long, and the use of a single
> +mutex for regular read/write and BRA is a show-stopper. Independent
> +operation of the control/command and BRA transfers is a fundamental
> +requirement, e.g. to change the volume level with the existing regmap
> +interface while downloading firmware.
> +
> +This places the burden on the codec driver to verify that there is no
> +concurrent access to the same address with the command/control
> +protocol and the BRA protocol.
> +
> +In addition, the 'sdw_msg' structure hard-codes support for 16-bit
> +addresses and paging registers which are irrelevant for BPT/BRA
> +support based on native 32-bit addresses. A separate API with
> +'sdw_bpt_msg' makes more sense.
> +
> +One possible strategy to speed-up all initialization tasks would be to
> +start a BRA transfer for firmware download, then deal with all the
> +"regular" read/writes in parallel with the command channel, and last
> +to wait for the BRA transfers to complete. This would allow for a
> +degree of overlap instead of a purely sequential solution. As a
> +results, the BRA API must support async transfers and expose a
> +separate wait function.

That sounds logical to me

> +
> +Error handling
> +~~~~~~~~~~~~~~
> +
> +The expected response to a 'bra_message' and follow-up behavior may
> +vary:
> +
> +   (1) A Peripheral driver may want to receive an immediate -EBUSY
> +       response if the BRA protocol is not available at a given time.
> +
> +   (2) A Peripheral driver may want to wait until a timeout for the
> +       on-going transfer to be handled
> +
> +   (3) A Peripheral driver may want to wait until existing BRA
> +       transfers complete or deal with BRA as a background task when
> +       audio transfers stop. In this case, there would be no timeout,
> +       and the operation may not happen if the platform is suspended.

Is this runtime suspend or S3/S4 case?

> +BTP/BRA API available to peripheral drivers
> +-------------------------------------------
> +
> +ASoC Peripheral drivers may use
> +
> +   - sdw_bpt_stream_open(mode)
> +
> +      This function verifies that the BPT protocol with the
> +      'mode'. For now only BRA is accepted as a mode. This function
> +      allocates a work buffer internally. This buffer is not exposed
> +      to the caller.
> +
> +     errors:
> +         -ENODEV: BPT/BRA is not supported by the Manager.
> +
> +         -EBUSY: another agent is already using the audio payload for
> +          audio transfers. There is no way to predict when the audio
> +          streams might stop, this will require the Peripheral driver
> +          to fall back to the regular (slow) command channel.
> +
> +         -EAGAIN: another agent is already transferring data using the
> +          BPT/BRA protocol. Since the transfers will typically last
> +          10s or 100s of ms, the Peripheral driver may wait and retry
> +          later.
> +
> +    - sdw_bpt_message_send_async(bpt_message)

why not have a single API that does both? First check if it is supported
and then allocate buffers and do the transfer.. What are the advantages
of using this two step process

-- 
~Vinod