Greg KH wrote:

>  
>
>>For the JSM driver, its up to you whether you feel its needed or not.
>>
>>However, I would like to mention that the DIGI drivers that currently
>>reside in the kernel sources *do* reserve that ioctl space,
>>and is acknowledged by "Documentation/ioctl-number.txt":
>>    
>>
>>>d'     F0-FF   linux/digi1.h
>>>      
>>>
>>I understand that the list is not a reservation list,
>>but a current list of potential ioctl conflicts...
>>    
>>
>
>It's not a reservation issue, it's the fact that we don't want to allow
>new ioctls, and if we do, they had better work properly (your
>implementation does not.)
>
>
>  
>
Hi Greg,

We understood that you don't want more new ioctls in the driver.  
For DPA tool(it is very good for customers) works, I think first I want 
to make this part code working, then you can decide if you pickup or not.

Old ioctls work with DPA tool. From your comments, looks old 
implementation does not work.
Need I use "register_ioctl32_convension" and 
"unregister_ioctl32_convension"?

Thank you so much!
wendy