Re: [PATCH v2 01/11] ASoC: topology: Able to create BE DAIs

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1.1: Type: text/plain, Size: 1480 bytes --]

On Fri, Sep 09, 2016 at 07:43:11PM +0800, mengdong.lin@linux.intel.com wrote:

> From: Mengdong Lin <mengdong.lin@linux.intel.com>

> Topology will check with ASoC if a BE DAI already exists by checking
> its name. If the BE DAI doesn't exist, topology will create a new one
> and the dai_load ops will be called for device specific init.

I still remain fundamentally unconvinced that this is a good or scalable
idea and that it makes sense to have it be part of the standard topology
ABI.  We are not going to download a DSP firmware that creates a new
physical link in the system so having the topology be the thing creating
those physical links seems wrong.

> create the HDA DAIs. Ideally, the driver should query BIOS to get the
> physical pin output types and connections to peripherals for a specific
> platform, but unfortunately BIOS sometimes does not have accurate info,
> so there is still hard code of BE DAI & DAI links in driver. To share a

Having the connections to the DSP specified in the toplogy is of course
sensible but fundamentally this is always going to be choosing from a
list of physical DAIs that we already know about and potential
connections for those that are also already known.  If the driver wants
to do something in response to a DAI being hooked up into the topology
(which seems to be the actual goal here) that would be sensible but
attempting to create physical hardware in a topology does not seem like
a good idea as a generic feature.

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]