Re: [PATCH 08/18] staging: wilc1000: remove declaration of WILC_WFI_frame_register

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Sep 21, 2015 at 03:20:12PM +0900, Tony Cho wrote:
> 
> 
> On 2015년 09월 21일 15:07, Greg KH wrote:
> >On Mon, Sep 21, 2015 at 02:54:28PM +0900, Tony Cho wrote:
> >>
> >>On 2015년 09월 21일 11:01, Greg KH wrote:
> >>>On Sun, Sep 20, 2015 at 03:51:15PM +0900, Chaehyun Lim wrote:
> >>>>This patch removes function declaration of WILC_WFI_frame_register in
> >>>>linux_wlan.c file, then adds it in wilc_wfi_cfgoperations.h file.
> >>>>
> >>>>The compilation warning occurs because it is assigned to incorrect
> >>>>pointer type of second parameter of WILC_WFI_frame_register.
> >>>>It is assigned with struct wireless_dev pointer type.
> >>>How was this code ever even working properly if the wrong pointer was
> >>>being passed into it?
> >>>
> >>>Tony, what is going on here, is this code just never called?  If so, can
> >>>we just delete the whole function?
> >>>
> >>>thanks,
> >>>
> >>>greg k-h
> >>This file with linux_wlan.c file is being fully changed, but at this time, the second parameter which passed as
> >>net_device pointer instead of wireless_dev incorrectly is never being used inside the WILC_WFI_frame_register().
> >>That's why no problem is found yet.
> >Then we should just delte the parameter entirely, no need to pass
> >something into a function that is never used.
> 
> This function is defined in the struct cfg80211_ops as the followings:
> 
> 	void	(*mgmt_frame_register)(struct wiphy *wiphy,
> 				
> 
>                                                 struct wireless_dev *wdev,
> 				
> 
>                                                 u16 frame_type, bool reg);
> 
> So, the second parameter cannot be removed to connect the cfg80211 structure correctly. The second parameter
> is not used just in wilc driver. It seems the second parameter is not famous in cfg80211 driver when looking for
> other cfg80211 driver, just to get the private data, but the wilc driver gets it from struct wiphy.

Ah, ok, that makes a bit more sense, thanks for explaining it, I'll go
try to queue up the patch now...

thanks,

greg k-h