On Mon, Jan 18, 2010 at 11:58:45AM -0600, Miguel Aguilar wrote:
Hi Mark,
- (*pdev)->dev.parent = davinci_vc->dev;
- platform_set_drvdata(*pdev, davinci_vc);
Newer drivers are tending to do this by looking at dev->parent in the child device rather than using the
Can you tell me what is driver that is using this new tend? Is this a wrong way to register the clients?
wm831x is one example, but there's others.
+#include <linux/delay.h> +#include <linux/pm.h> +#include <linux/i2c.h>
Not needed.
What headers are not needed?
linux/i2c.h
snd_soc_cnew(&cq93vc_snd_controls[i],codec, NULL));if (err < 0)return err;- }
This is snd_soc_add_controls().
So I can call directly snd_soc_add_controls instead of using the function above?
Yes, your code above is just an open coded version of that function so you can just replace your function with a call to snd_soc_add_controls().
These are actually mostly redundant with this hardware - the core will bring the driver down to _STANDBY before suspending the driver and your _STANDBY and _OFF states are equivalent. This means that both functions should end up being noops. On the other hand they do no harm.
This means that I can get rid of the function suspend?
Yup.
- /* Set the DAC digital attenuation to 0 dB */
- cq93vc_write(codec, DAVINCI_VC_REG09, DAVINCI_VC_REG09_DIG_ATTEN);
The standard thing for things like this is to leave the defaults in the driver as whatever the hardware default is then let either the machine driver or (better) userspace override it. This avoids issues with defaults being good for one system and not another.
So, Should I remove the values that I seeting above, and let the hardware use its default values?
Yes.
This has been removed in favour of relying on more generic kernel mechanisms (hopefully).\
What was exactly removed here the NULL version of the functions or the functions itself.
The functions themselves.