Is platform_device_register_simple() deprecated?

Is platform_device_register_simple() deprecated?

[Rene Herman]

Hi Greg, Russel, Dmitry.

ALSA is using platform_device_register_simple(). Jean Delvare pointed:

http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2

out, where _simple looks to be slated for removal. Is this indeed the 
case? ALSA isn't using the resources -- doing a manual alloc/add would 
not be a problem...

Rene.

Re: Is platform_device_register_simple() deprecated?

[Russell King]

On Wed, Apr 12, 2006 at 07:50:37PM +0200, Rene Herman wrote:
> Hi Greg, Russel, Dmitry.
> 
> ALSA is using platform_device_register_simple(). Jean Delvare pointed:
> 
> http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
> 
> out, where _simple looks to be slated for removal. Is this indeed the 
> case? ALSA isn't using the resources -- doing a manual alloc/add would 
> not be a problem...

It would be good to remove the old (restricted) API - I'm of the opinion
that having too many interfaces just adds extra maintainence burden, so
we should strive to ensure that old APIs are deprecated and removed in
a timely fashion.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

Re: Is platform_device_register_simple() deprecated?

[Greg KH]

On Wed, Apr 12, 2006 at 07:50:37PM +0200, Rene Herman wrote:
> Hi Greg, Russel, Dmitry.
> 
> ALSA is using platform_device_register_simple(). Jean Delvare pointed:
> 
> http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
> 
> out, where _simple looks to be slated for removal. Is this indeed the 
> case? ALSA isn't using the resources -- doing a manual alloc/add would 
> not be a problem...

Great, care to convert ALSA to use the proper api so we can remove
platform_device_register_simple()?

thanks,

greg k-h

Re: Is platform_device_register_simple() deprecated?

[Kumar Gala]

On Apr 12, 2006, at 4:41 PM, Greg KH wrote:

> On Wed, Apr 12, 2006 at 07:50:37PM +0200, Rene Herman wrote:
>> Hi Greg, Russel, Dmitry.
>>
>> ALSA is using platform_device_register_simple(). Jean Delvare  
>> pointed:
>>
>> http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
>>
>> out, where _simple looks to be slated for removal. Is this indeed the
>> case? ALSA isn't using the resources -- doing a manual alloc/add  
>> would
>> not be a problem...
>
> Great, care to convert ALSA to use the proper api so we can remove
> platform_device_register_simple()?

Can we mark this deprecated and add it to feature-removal-schedule.txt.

I'm using it about half a dozen times in arch/powerpc and wasn't sure  
if it was planned on being removed.

- kumar

Re: Is platform_device_register_simple() deprecated?

[Greg KH]

On Wed, Apr 12, 2006 at 06:09:48PM -0500, Kumar Gala wrote:
> 
> On Apr 12, 2006, at 4:41 PM, Greg KH wrote:
> 
> >On Wed, Apr 12, 2006 at 07:50:37PM +0200, Rene Herman wrote:
> >>Hi Greg, Russel, Dmitry.
> >>
> >>ALSA is using platform_device_register_simple(). Jean Delvare  
> >>pointed:
> >>
> >>http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
> >>
> >>out, where _simple looks to be slated for removal. Is this indeed the
> >>case? ALSA isn't using the resources -- doing a manual alloc/add  
> >>would
> >>not be a problem...
> >
> >Great, care to convert ALSA to use the proper api so we can remove
> >platform_device_register_simple()?
> 
> Can we mark this deprecated and add it to feature-removal-schedule.txt.

Sure, I'll take a patch for that.  But really, it's just easier to fix
up all callers and delete the function.  It isn't anything that
feature-removal-schedule.txt should care about, as it's just the normal
API changes we do all the time.

thanks,

greg k-h

Re: Is platform_device_register_simple() deprecated?

[Rene Herman]

Greg KH wrote:

>> ALSA is using platform_device_register_simple(). Jean Delvare pointed:
>>
>> http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
>>
>> out, where _simple looks to be slated for removal. Is this indeed the 
>> case? ALSA isn't using the resources -- doing a manual alloc/add would 
>> not be a problem...
> 
> Great, care to convert ALSA to use the proper api so we can remove
> platform_device_register_simple()?

Sure. Before I go over them though, could you perhaps confirm that just 
doing a manual alloc/add _is_ this proper API? Ie, something like:

     device = platform_device_alloc(NAME, i);
     if (!device)
             return -ENOMEM;

     error = platform_device_add(device);
     if (error) {
             platform_device_put(device);
             return error;
     }

(there by the way are still a few users left outside ALSA as well)

Rene.

Re: Is platform_device_register_simple() deprecated?

[Kumar Gala]

On Apr 12, 2006, at 6:30 PM, Greg KH wrote:

> On Wed, Apr 12, 2006 at 06:09:48PM -0500, Kumar Gala wrote:
>>
>> On Apr 12, 2006, at 4:41 PM, Greg KH wrote:
>>
>>> On Wed, Apr 12, 2006 at 07:50:37PM +0200, Rene Herman wrote:
>>>> Hi Greg, Russel, Dmitry.
>>>>
>>>> ALSA is using platform_device_register_simple(). Jean Delvare
>>>> pointed:
>>>>
>>>> http://marc.theaimsgroup.com/?l=linux-kernel&m=113398060508534&w=2
>>>>
>>>> out, where _simple looks to be slated for removal. Is this  
>>>> indeed the
>>>> case? ALSA isn't using the resources -- doing a manual alloc/add
>>>> would
>>>> not be a problem...
>>>
>>> Great, care to convert ALSA to use the proper api so we can remove
>>> platform_device_register_simple()?
>>
>> Can we mark this deprecated and add it to feature-removal- 
>> schedule.txt.
>
> Sure, I'll take a patch for that.  But really, it's just easier to fix
> up all callers and delete the function.  It isn't anything that
> feature-removal-schedule.txt should care about, as it's just the  
> normal
> API changes we do all the time.


Well, it is an exported interface so I figured that fit in the same  
category as removing the export of insert_resource.

But, I'm not too concerned one way or the other about it.

- k