spi codec names in machine drivers

spi codec names in machine drivers

[Mike Frysinger]

after the recent multi-component patch, spi codec names were changed
to "<name>.<spics>" in the machine drivers.  so if we have an ad193x
codec on spi bus 0 cs 5, the name is now "ad193x.5".

this also seems to be how fmt_single_name in soc-core.c is doing
things.  am i reading this right ?

what if i have two codecs on two different spi busses but happen to
have the same cs ?  shouldnt the spi code do the same as i2c and
include the bus in the id naming ?
-mike

Re: spi codec names in machine drivers

[Mark Brown]

On Tue, Mar 29, 2011 at 02:05:04AM -0400, Mike Frysinger wrote:
> after the recent multi-component patch, spi codec names were changed
> to "<name>.<spics>" in the machine drivers.  so if we have an ad193x
> codec on spi bus 0 cs 5, the name is now "ad193x.5".
> 
> this also seems to be how fmt_single_name in soc-core.c is doing
> things.  am i reading this right ?
> 
> what if i have two codecs on two different spi busses but happen to
> have the same cs ?  shouldnt the spi code do the same as i2c and
> include the bus in the id naming ?

I think we should just switch to using dev_name() everywhere.

Re: spi codec names in machine drivers

[Mike Frysinger]

On Wed, Mar 30, 2011 at 17:20, Mark Brown wrote:
> On Tue, Mar 29, 2011 at 02:05:04AM -0400, Mike Frysinger wrote:
>> after the recent multi-component patch, spi codec names were changed
>> to "<name>.<spics>" in the machine drivers.  so if we have an ad193x
>> codec on spi bus 0 cs 5, the name is now "ad193x.5".
>>
>> this also seems to be how fmt_single_name in soc-core.c is doing
>> things.  am i reading this right ?
>>
>> what if i have two codecs on two different spi busses but happen to
>> have the same cs ?  shouldnt the spi code do the same as i2c and
>> include the bus in the id naming ?
>
> I think we should just switch to using dev_name() everywhere.

so you'd want the machine driver to use "spi0.5" ?  or "ad193x-spi0.5" ?
-mike
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: spi codec names in machine drivers

[Mark Brown]

On Wed, Mar 30, 2011 at 05:58:29PM -0400, Mike Frysinger wrote:

> so you'd want the machine driver to use "spi0.5" ?  or "ad193x-spi0.5" ?

The latter, I guess.

Re: spi codec names in machine drivers

[Mike Frysinger]

On Wed, Mar 30, 2011 at 20:11, Mark Brown wrote:
> On Wed, Mar 30, 2011 at 05:58:29PM -0400, Mike Frysinger wrote:
>> so you'd want the machine driver to use "spi0.5" ?  or "ad193x-spi0.5" ?
>
> The latter, I guess.

makes sense to me

but back to the original question, how is it working today ? :)  i
want to make sure the current Blackfin machine drivers are written
correctly, and the boards i readily have access to are either AC97 or
I2C, so it's difficult for me to boot up and verify myself.
-mike
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: spi codec names in machine drivers

[Mark Brown]

On Wed, Mar 30, 2011 at 08:15:14PM -0400, Mike Frysinger wrote:

> but back to the original question, how is it working today ? :)  i
> want to make sure the current Blackfin machine drivers are written
> correctly, and the boards i readily have access to are either AC97 or
> I2C, so it's difficult for me to boot up and verify myself.

This is the common problem with SPI codecs, they're very rarely used so
not frequently tested.

Re: spi codec names in machine drivers

[Mike Frysinger]

On Wed, Mar 30, 2011 at 20:30, Mark Brown wrote:
> On Wed, Mar 30, 2011 at 08:15:14PM -0400, Mike Frysinger wrote:
>> but back to the original question, how is it working today ? :)  i
>> want to make sure the current Blackfin machine drivers are written
>> correctly, and the boards i readily have access to are either AC97 or
>> I2C, so it's difficult for me to boot up and verify myself.
>
> This is the common problem with SPI codecs, they're very rarely used so
> not frequently tested.

so i need to scrounge up some addon boards and figure out the answer myself ;)

hmm, i wonder if it'd be easier to finish my SPI simulator and write a
small simulated audio codec device to connect to it ...
-mike
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: spi codec names in machine drivers

[Mike Frysinger]

On Wed, Mar 30, 2011 at 20:32, Mike Frysinger wrote:
> On Wed, Mar 30, 2011 at 20:30, Mark Brown wrote:
>> On Wed, Mar 30, 2011 at 08:15:14PM -0400, Mike Frysinger wrote:
>>> but back to the original question, how is it working today ? :)  i
>>> want to make sure the current Blackfin machine drivers are written
>>> correctly, and the boards i readily have access to are either AC97 or
>>> I2C, so it's difficult for me to boot up and verify myself.
>>
>> This is the common problem with SPI codecs, they're very rarely used so
>> not frequently tested.
>
> so i need to scrounge up some addon boards and figure out the answer myself ;)

for the curious, the answer is as i feared ... the codec_name needed
in the machine driver is "spi<busnum>.<csnum>".  the codec name should
not be in that string.  unlike the i2c version which should be
"<codec>.<busnum>-<id>".
-mike
_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel