On Mon, Jan 19, 2015 at 7:28 PM, Lars-Peter Clausen lars@metafoo.de wrote:

On 01/19/2015 04:46 PM, Manuel Lauss wrote:

...

Hi Lars,

Since commit 6794f709b7124ff1e574c4f4c9494418ab56c4b4 (ASoC: ac97: Drop delayed device registration) the wm9712 ts on my MIPS DB1300 board doesn't work any more:

wm97xx-ts 0-0:wm9712-codec.1: Device with vendor 0000 is not a wm97xx db1200-ac97 db1300-ac97.0: wm9712-hifi <-> au1xpsc_ac97.1 mapping ok wm9712-codec wm9712-codec.1: ASoC: mux Differential Source has no paths wm9712-codec wm9712-codec.1: ASoC: mux Capture Phone Mux has no paths db1200-ac97 db1300-i2s.0: wm8731-hifi <-> au1xpsc_i2s.2 mapping ok

whereas before this patch: [ 1.820000] db1200-ac97 db1300-ac97.0: wm9712-hifi <-> au1xpsc_ac97.1 mapping ok [ 1.820000] wm9712-codec wm9712-codec.1: ASoC: mux Differential Source has no paths [ 1.830000] wm9712-codec wm9712-codec.1: ASoC: mux Capture Phone Mux has no paths [ 1.840000] wm97xx-ts 0-0:wm9712-codec.1: detected a wm9712 codec [ 1.850000] input: wm97xx touchscreen as /devices/platform/db1300-ac97.0/0-0:wm9712-codec.1/input/input0 [ 1.860000] db1200-ac97 db1300-i2s.0: wm8731-hifi <-> au1xpsc_i2s.2 mapping ok

Do you have any ideas how to solve this?

For a proper fix probably rewrite the AC97 bus support. But can you try if the attached patch works as a workaround?

Yes it does.

Thank you! Manuel

On Mon, Jan 19, 2015 at 07:58:22PM +0100, Lars-Peter Clausen wrote:

Ok, thanks for testing. I'll try to put together a proper patch and then send it out tomorrow or so. The wm9705 and wm9713 look like they need the same patch. So maybe it is possible to add support for reseting the device in the core itself rather than doing it for every device in the device driver.

There was the ucb1400 (or some similar part number too) which was the same deal with an integrated touchscreen and CODEC and probably needs the same thing - it used ac97.c.