[alsa-devel] [PATCH v4 0/7] ASoC: intel - add skylake PCM driver

[Takashi Iwai]

At Mon, 25 May 2015 16:54:58 +0530,
Vinod Koul wrote:
> 
> On Mon, May 25, 2015 at 08:57:05AM +0200, Takashi Iwai wrote:
> > At Mon, 11 May 2015 16:23:58 +0530,
> > Vinod Koul wrote:
> > > 
> > > SKL has HDA controller based audio subsystem with DSP and support for I2S,
> > > HDA, PDM links. The hda core code has been moved to sound/hda/ by Takashi
> > > which current HDA drivers use and will also be used by ASoC SKL driver.
> > > 
> > > The SKL platform driver will load and create the soc_hdac_bus which embeds
> > > the hdac_bus, same for hdac_device (hda codecs) and hdac_stream (pcms) This
> > > is on top of hdac code in Takashi's topic/hda
> > > 
> > > This patch provides the match function for asoc type hda codecs and let's
> > > them get enumerated by hdac. The second patch in this series adds the
> > > controller specific soc code. Common parts are in hdac core with changes
> > > introduced as part of SKL controller in soc part. Then we add the rest of
> > > controller PCM driver code (still HDA) and last patch breaks the HDA streams
> > > to host and link which will allow insertion of DSP in between these links.
> > > 
> > > The subsequent series will add IPC driver for SKL (using common IPC
> > > routines), then DSP topology handlers, DSP code with I2S support and then
> > > lastly when DFW is accepted then its handlers.
> > > 
> > > This patch series adds the hda codec match functions followed by asoc hda
> > > controller routines, then SKL PCM driver and last decouples the controller
> > > for splitting the links
> > 
> > Merging this without the codec driver would be rather confusing for
> > users, since it gives a conflicting driver that doesn't work at all.
> > I suppose that the merge should be pending until the complete
> > implementation of ASoC HDA.
> Well users wont notice till we have a machine driver which selects this
> driver. That is why machine driver will come last and needs to be
> merged only after controller and codec drivers are merged for end
> users to notice

We still don't provide a way to lead to the right driver.  With your
patch 5, it creates another SKL PCI driver that conflicts with the
existing driver entry of snd-hda-intel.  So, at this moment, it just
give a conflict for non-working driver, and it's not good to leave so
unless the complete driver set is provided in a week or so.


Takashi