mbox series

[v3,0/6] ASoC: extra format on each DAI

Message ID 87zfk4o5l2.wl-kuninori.morimoto.gx@renesas.com (mailing list archive)
Headers show
Series ASoC: extra format on each DAI | expand

Message

Kuninori Morimoto Jan. 6, 2025, 5:48 a.m. UTC
Hi Mark
Cc Stephen

Current clock provider/consumer setting is set by dai_link->dai_fmt, and it
is Codec base on Sound Card driver (= SND_SOC_DAIFMT_CBx_CFx).

Current CPU/Codec drivers are already based on its own base
(= SND_SOC_DAIFMT_Bx_Fx). So, Codec clock setting uses dai_link->dai_fmt
as-is, and CPU side clock setting is created from Codec base setting by
flipping. Because of this, we can't set both CPU/Codec clock consumer for
example.

To solve this issue, this patch-set adds new ext_fmt on each DAI.
It can keep compatible with legacy style.

	1. SND_SOC_DAIFMT_FORMAT_MASK
	2. SND_SOC_DAIFMT_CLOCK
	3. SND_SOC_DAIFMT_INV
	4. SND_SOC_DAIFMT_CLOCK_PROVIDER

	dai_fmt : dai_link->dai_fmt = common settings
	ext_fmt : each DAI settings
    
Legacy
	dai_fmt  includes 1, 2, 3, 4
    
New style
	dai_fmt  includes 1, 2, 3
	ext_fmt  includes 4

Audio-Graph-Card2 will use this new style by this patch-set.
By this patch, Card2 default behavior (= no "clock-master / frame-master"
settings on DT) will be changed, but no drivers are using it.

In case of no DAI has "clock-master / frame-master" property on DT,
it will be...

Legacy
	CPU  : provider (because flipped from Codec)
	Codec: consumer

New style
	CPU  : consumer
	Codec: consumer

One note is that Simple-Card, Audio-Graph-Card don't implement
this new style to keep compatiblily.

In Overlay case, port order can be random, so we shouldn't use get_next()
function to get next port.

[1/7] - [4/7] are just cleanup before new feature.
[5/7]         adjust to overlay
[6/7] - [7/7] adds new daifmt

This patch-set needs Stephen's Tested-by

v2 -> v3
	- use of_graph_get_port_by_id() instead of of_graph_get_next_port()

v1 -> v2
	- based on asoc/for-next to avoid conflict issue.
	- Needs Stephen's Tested-by

Link: https://lore.kernel.org/r/87r064cxym.wl-kuninori.morimoto.gx@renesas.com
Link: https://lore.kernel.org/r/8734innkpy.wl-kuninori.morimoto.gx@renesas.com

Kuninori Morimoto (7):
      ASoC: audio-graph-card2: use __free(device_node) for device node
      ASoC: audio-graph-card: use __free(device_node) for device node
      ASoC: simple-card: use __free(device_node) for device node
      ASoC: soc-core: return 0 if np was NULL on snd_soc_daifmt_parse_clock_provider_raw()
      ASoC: audio-graph-card2: use of_graph_get_port_by_id() at graph_get_next_multi_ep()
      ASoC: soc-core: Enable to use extra format on each DAI
      ASoC: audio-graph-card2: Use extra format on each DAI

Comments

Stephen Gordon Jan. 6, 2025, 8:53 a.m. UTC | #1
On 6/01/2025 4:48 pm, Kuninori Morimoto wrote:

> This patch-set needs Stephen's Tested-by
> 
> Kuninori Morimoto (7):
>        ASoC: audio-graph-card2: use __free(device_node) for device node
>        ASoC: audio-graph-card: use __free(device_node) for device node
>        ASoC: simple-card: use __free(device_node) for device node
>        ASoC: soc-core: return 0 if np was NULL on snd_soc_daifmt_parse_clock_provider_raw()
>        ASoC: audio-graph-card2: use of_graph_get_port_by_id() at graph_get_next_multi_ep()
>        ASoC: soc-core: Enable to use extra format on each DAI
>        ASoC: audio-graph-card2: Use extra format on each DAI
> 

All patches:

Tested-by: Stephen Gordon <gordoste@iinet.net.au>


Thanks

Stephen