[PATCH] ASoC: core: Fix snd_soc_add_card_controls to pass correct data

[PATCH] ASoC: core: Fix snd_soc_add_card_controls to pass correct data

[Stephen Warren]

Commit 022658b "ASoC: core: Add support for DAI and machine kcontrols."
created new functions to create controls for cards, codecs, platforms,
and DAIs. The new snd_soc_add_card_controls() passed the wrong "data" to
utility function snd_soc_add_controls(); it should be passing a codec,
but instead ended up passing the soc_card. Fix this.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
---
 sound/soc/soc-core.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/sound/soc/soc-core.c b/sound/soc/soc-core.c
index 8bd9995..e44b6f6 100644
--- a/sound/soc/soc-core.c
+++ b/sound/soc/soc-core.c
@@ -2095,9 +2095,13 @@ int snd_soc_add_card_controls(struct snd_soc_card *soc_card,
 	const struct snd_kcontrol_new *controls, int num_controls)
 {
 	struct snd_card *card = soc_card->snd_card;
+	struct snd_soc_codec *codec =
+			list_first_entry(&soc_card->codec_dev_list,
+					 struct snd_soc_codec,
+					 card_list);
 
 	return snd_soc_add_controls(card, soc_card->dev, controls, num_controls,
-			NULL, soc_card);
+			NULL, codec);
 }
 EXPORT_SYMBOL_GPL(snd_soc_add_card_controls);
 
-- 
1.7.0.4

Re: [PATCH] ASoC: core: Fix snd_soc_add_card_controls to pass correct data

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 976 bytes --]

On Tue, Feb 14, 2012 at 03:56:33PM -0700, Stephen Warren wrote:
> Commit 022658b "ASoC: core: Add support for DAI and machine kcontrols."
> created new functions to create controls for cards, codecs, platforms,
> and DAIs. The new snd_soc_add_card_controls() passed the wrong "data" to
> utility function snd_soc_add_controls(); it should be passing a codec,
> but instead ended up passing the soc_card. Fix this.

Well, sort of.  The idea with Liam's commit was to split things out so
that card operations got given a card so the argument change was
intentional.  What's gone wrong is that things like PIN_SWITCH() (which
is only ever used by cards) weren't updated to match so they still
expect to find a CODEC there.  This'll quite probably have broken some
other machine specific controls too.

I'll send a patch for PIN_SWITCH later today but it looks like we need
an audit of all the drivers using non-CODEC adds to make sure they don't
assume a CODEC in the callbacks.

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]