[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <f4a71a23-a3ca-42ec-97ee-03e70e369db4@sirena.org.uk>
Date: Wed, 30 Jul 2025 13:46:24 +0100
From: Mark Brown <broonie@...nel.org>
To: ew kim <ew.kim@...sung.com>
Cc: s.nawrocki@...sung.com, robh@...nel.org, krzk+dt@...nel.org,
lgirdwood@...il.com, tiwai@...e.com, perex@...ex.cz,
conor+dt@...nel.org, alim.akhtar@...sung.com,
linux-sound@...r.kernel.org, devicetree@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org,
linux-samsung-soc@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: Add ExynosAuto ABOX generic platform and PCM support
On Mon, Jul 21, 2025 at 11:30:43AM +0900, ew kim wrote:
> This patch series adds the ABOX Generic audio management driver for
> Samsung ExynosAuto SoCs, along with IPC messaging and PCM frontend
> drivers, including their corresponding device tree bindings.
>
> ### ABOX Architecture Design: Fixed and Variable Structure
>
> The ABOX audio framework is designed with a clear split between:
> - **Fixed part** (common framework): reusable components across SoCs
> - Generic ABOX platform driver
> - IPC messaging handler
> - PCM frontend with ALSA compressed audio support
> - Solution manager for dynamic control
> - **Variable part** (SoC-specific): defined via device tree
> - IRQ numbers, stream IDs, buffer sizes, and ADSP allocation
> - Defined per PCM/IPC device node in DTS
This all sounds from a system integration point of view like a fairly
standard audio DSP. Usually something like that would be structured
with the generic bits of DSP support done as a library which the drivers
for specific bits of hardware link to, the SOF code is the most obvious
example of this we have upstream but there's a few other simpler ones
like the Cirrus CODECs. If there's a reason why this wouldn't work here
it's not clear to me from what you've posted thus far.
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists