Re: [PATCH RFC 0/3] staging: r8188eu: avoid uninit value bugs

[Greg KH]

On Sun, Aug 22, 2021 at 03:50:35PM +0300, Pavel Skripkin wrote:
> On 8/22/21 3:39 PM, Greg KH wrote:
> > > 
> > > Yes, but _rtw_read*() == 0 indicates 2 states:
> > > 
> > > 	1. Error on transfer side
> > > 	2. Actual register value is 0
> > 
> > That's not a good design, it should be fixed.  Note there is the new
> > usb_control_msg_recv() function which should probably be used instead
> > here, to prevent this problem from happening.
> > 
> 
> Thank you, Greg, for confirmation. That's was the point why I started to
> write this series :)
> 
> I think, usb_control_msg_recv() won't help us with this problem, since all
> rtw_read*() functions return an unsigned value now. In future, when driver
> code will be fixed (ex: a lot of void function, which can fail and leave
> passed pointer uninitialized) we can move to new usb API and then move
> driver out of staging :)

That function _should_ be used at the lower levels of this call chain.
If you want to go from the top -> down or bottom -> up of fixing this is
your choice, but as others have pointed out, this patch series as-is
still needs some work.

thanks,

greg k-h