On Sat, May 29, 2021 at 02:13:45PM +0200, Martin Kaiser wrote: > Remove the wext handlers that link to an empty dummy function. > > Signed-off-by: Martin Kaiser <martin@xxxxxxxxx> > --- > If no handler is installed for a wext ioctl, the wext core will return > -ENOTSUPP to user space. The dummy function returned -1. It could be argued > that this change breaks the user space ABI. > > However, it's rather unlikely that an application expects a particular wlan > driver and chipset when it configures a wireless network. Checking for > errno==1 explicitly will already be non-portable as most other drivers set > errn==ENOTSUPP for non-existing ioctls. > The patch is good, but next time, just put all this commentary in the commit message. It has to do with the impact of the patch on userspace and it's helpful for the reviewers. Some reviewers won't need it, but I try to target my commit messages at developers who aren't subsystem experts. I also like that it shows you thought about it. I recently saw someone claim that "anyone who can review these patches will already know what the <SOMETHING> acronym stands for so that's why I didn't explain." And I was like, "Oh wow... That's a harsh burn on me!" regards, dan carpenter
