[Letux-kernel] LetuxOS: Kernel: m-rc 5.3-rc1 successful

[H. Nikolaus Schaller]

Some additional thoughts:

* here is the "Issue" record: http://projects.goldelico.com/p/gta04-kernel/issues/883/

* the letux/pyra-minimal branch is without any AESS patch (on top of 5.3-rc2).
  I tried audio yesterday on the Pyra but did not get more than scratchy noise :(

* There was a really good description on omappedia but I think it has gone.

* I think the Droid 4 is omap4430 + twl6040 + telephony modem and therefore the
developers may also be interested in this.


> Am 30.07.2019 um 17:21 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
> 
> Hi,
> 
>> Am 30.07.2019 um 16:29 schrieb N. Jackson <nljlistbox2 at gmail.com>:
>> 
>> At 10:22 +0200 on Tuesday 2019-07-23, H. Nikolaus Schaller wrote:
>>> 
>>> a) the AESS patches by Peter we are carrying along did no
>>> longer compile.
>>> 
>>> Reason is that some members of the dai_link struct have
>>> been replaced by pointers to other structs. The macros
>>> used to initialize them can no longer be used.
>>> 
>>> Therefore I have simply disabled the offending code
>>> lines.
>>> 
>>> Note: this is not yet fixed in the letux/pyra-minimal tree.
>>> 
>>> Generally we need someone who looks at this code pile
>>> and brings them into a better shape. Ideally it would
>>> be made working incl. AESS firmware load and pushed
>>> upstream :)
>> 
>> What is the rationale for the AESS patches in Letux? Does Linux
>> currently have no AESS support, or do these patches modify
>> existing Linux AESS support? If the latter, are the AESS patches
>> in Letux fixing bugs or adding features?
> 
> No, there is no AESS in kernel.org. Andrej did port an older
> version written by Peter (CC:) and I made it work (well compile)
> again on newer mainline kernels.
> 
> What is not working is firmware loading.
> 
> And there are bugs in the integration (API to setup DAI links etc.
> has changed).
> 
> So we carry that code along in the hope that Peter or someone
> else finds the time to fix the issues introduced by not having
> it upstream.
> 
>> 
>> If the former (no existing AESS support in Linux), then does the
>> hardware targeted by Letux not have audio capabilities
>> (alternative to AESS) for which there is existing support in
>> Linux?
> 
> Yes. Of course. Think of it as a coprocessor like a GPU.
> 
> Without AESS we can use the upstream twl6040 driver directly
> and use it as a simple sound card through aplay. This is the
> status you get if you use omap2plus_defconfig and/or build a
> kernel for Pandaboard or OMAP5EVM.
> 
> But AESS has some nice features:
> * can run independently of OMAP5
> * allows to mix headset, modem audio, bluetooth microphone etc.
>  i.e. between all McBSP interfaces
> * audio filters
> 
> These features make it interesting to use. But: this needs a
> binary blob (like GPU or WiFi chips), which is a no-go for FLOSS
> purists.
> 
> BTW: the OMAP4 + twl6040 (PandaBoard) has the same AESS.
> 
> BR,
> Nikolaus
> 
> _______________________________________________
> http://projects.goldelico.com/p/gta04-kernel/
> Letux-kernel mailing list
> Letux-kernel at openphoenux.org
> http://lists.goldelico.com/mailman/listinfo.cgi/letux-kernel