[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <599701D2.5080001@rock-chips.com>
Date: Fri, 18 Aug 2017 23:03:46 +0800
From: jeffy <jeffy.chen@...k-chips.com>
To: Mark Brown <broonie@...nel.org>
CC: linux-kernel@...r.kernel.org, dgreid@...omium.org, heiko@...ech.de,
briannorris@...omium.org, mka@...omium.org, dianders@...omium.org,
Jaroslav Kysela <perex@...ex.cz>, alsa-devel@...a-project.org,
Oder Chiou <oder_chiou@...ltek.com>,
Takashi Iwai <tiwai@...e.com>,
Liam Girdwood <lgirdwood@...il.com>,
Bard Liao <bardliao@...ltek.com>
Subject: Re: [PATCH v4 1/9] ASoC: rt5514: Avoid legacy dai naming
Hi Mark,
On 08/18/2017 07:45 PM, Mark Brown wrote:
> On Fri, Aug 18, 2017 at 11:11:39AM +0800, Jeffy Chen wrote:
>> Currently we are using devm_snd_soc_register_component, which would
>> use legacy dai naming when dai drv id is zero.
>>
>> Set a non-zero dai drv id to use dai drv name for dai name.
>
> Why? This is clearly not good, we shouldn't be expecting all devices
> with a single DAI to set the DAI number to 1 and the changelog doesn't
> articulate any reason why anything cares what the name is.
>
i was trying to avoid legacy dai naming, because snd_soc_find_dai will
try to match the dai name:
struct snd_soc_dai *snd_soc_find_dai(
const struct snd_soc_dai_link_component *dlc)
{
...
list_for_each_entry(dai, &component->dai_list, list) {
if (dlc->dai_name && strcmp(dai->name,
dlc->dai_name))
continue;
return dai;
if match failed, it would break soc_bind_dai_link:
static int soc_bind_dai_link(struct snd_soc_card *card,
struct snd_soc_dai_link *dai_link)
{
...
/* Find CODEC from registered CODECs */
codec_dais = rtd->codec_dais;
for (i = 0; i < rtd->num_codecs; i++) {
codec_dais[i] = snd_soc_find_dai(&codecs[i]);
if (!codec_dais[i]) {
dev_err(card->dev, "ASoC: CODEC DAI %s not
registered\n",
codecs[i].dai_name);
goto _err_defer;
}
when using legacy dai naming, the dai->name for rt5514-spi would be the
dev name, which is spi2.0 with my local 4.4 kernel, and would be
spi32765.0 with upstream kernel.
Powered by blists - more mailing lists