[MinnowBoard] Where is the MBM I2S driver in Linux

[li yuqian]

Hi Liam,

2015-06-04 23:07 GMT+08:00 Liam Girdwood <liam.r.girdwood at linux.intel.com>:

> On Thu, 2015-06-04 at 14:10 +0800, li yuqian wrote:
> > Hi Liam,
> >
> > 2015-06-03 16:55 GMT+08:00 Liam Girdwood
> > <liam.r.girdwood at linux.intel.com>:
> >         On Tue, 2015-06-02 at 18:27 +0100, Strasser, Kevin wrote:
> >         > Hi Li,
> >         >
> >         > On 06/01/15 8:03 PM, li yuqian wrote:
> >         > > check those driver, and i can understand there are three
> >         parts for the LPE
> >         > > driver to working, special for the LPE firmware need to
> >         adde it. i can figure
> >         > > it out in the
> >         > >
> >
> https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/tree/sound/soc/intel/sst-acpi.c?id=refs/tags/v4.0.4
> >         > > line 249 and 250.
> >         > >
> >         > > but when i checked that sst-acpi.c, i found the baytrial
> >         sst is not support
> >         > > the DMA data transfer, see the resindex_dma_base = -1 in
> >         line 262 and line 51
> >         > > /* DMA only valid when resindex_dma_base != -1*/
> >         > >
> >         > > Am i correct? the MBM i2s does not support DMA data
> >         transfer?
> >         >
> >         > You are correct, the Bay Trail LPE driver does not use DMA.
> >         I think that is just
> >         > a driver limitation, and does not necessarily indicate that
> >         the hardware isn't
> >         > capable.
> >         >
> >         > Liam, can you provide any context here?
> >
> >
> >         The BYT audio DSP driver does not use DMA for loading the FW
> >         from host
> >         memory to DSP memory, but it does use DMA for audio playback
> >         and capture
> >         (controlled by the FW). The BYT audio driver just does a
> >         memcpy() of the
> >         FW from host to DSP memory.
> >
> >         Liam
> >
> >         >
> >         > Thanks,
> >         > Kevin
> >
> >
> > Thanks for the explanation. is there more details information for that
> > "controlled by FW"? such as what is the data width of FW required, is
> > that fixed or user can change? in other words, should i use different
> > that FW to do different data width, sync clock and sst clock? also, i
> > found the TIZEN have some different FW for download in INTEL website
> > in
> >
> https://downloadcenter.intel.com/download/23903/Intel-Atom-Processor-E3800-Series-Linux-support-package-LPE-Firmware-for-Tizen-IVI-3-0-M2-Kernel-3-10-LTSI,
> can i use those FW to load to mainline kernel to support different mode of
> LPE, such as the master/slave mode with different sync, main clock?
> >
>
> It sounds like you have some different requirements to those supported
> by the FW ? Can you be more specific ? The current FW can support 16bit
> stereo at 48kHz with the codec as I2S slave.
>
> The FW wholly controls operation and configuration of the audio DMA and
> SSP ports (the host does not touch these). The host allocates the DMA
> buffers though.
>
> Btw, there are some new FW releases that are coming out soon that will
> provide some more flexibility wrt SSP configuration. This may be able to
> help with your requirements too.
>
> Thanks
>
> Liam
> >
> > thanks
>
>
>
Thank you. now i understand how is FW working. i project required the SSP
working at 8KHZ sync, 2M main TDM bus clock, slave mode, will INTEL release
such configure FW? i am expecting those new FW release ASAP ;)

Thanks
Yuqian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.elinux.org/pipermail/elinux-minnowboard/attachments/20150605/2d8a74ba/attachment.html>