Re: [PATCH] debugfs: remove return value of debugfs_create_bool()

[Greg Kroah-Hartman]

On Tue, May 25, 2021 at 09:26:42AM +0200, Geert Uytterhoeven wrote:
> Hi Greg,
> 
> On Mon, May 24, 2021 at 2:39 PM Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > On Mon, May 24, 2021 at 01:44:38PM +0200, Geert Uytterhoeven wrote:
> > > On Mon, May 24, 2021 at 12:18 PM Greg Kroah-Hartman
> > > <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > > On Mon, May 24, 2021 at 11:51:42AM +0200, Geert Uytterhoeven wrote:
> > > > > On Mon, May 24, 2021 at 11:41 AM Greg Kroah-Hartman
> > > > > <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > > > > On Mon, May 24, 2021 at 11:11:32AM +0200, Geert Uytterhoeven wrote:
> > > > > > > On Fri, May 21, 2021 at 10:28 PM Greg Kroah-Hartman
> > > > > > > <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > > > > > > > No one checks the return value of debugfs_create_bool(), as it's not
> > > > > > > > needed, so make the return value void, so that no one tries to do so in
> > > > > > >
> > > > > > > Please explain in the patch description why it is not needed.
> > > > > >
> > > > > > Because you just do not need it, like almost all other debugfs calls
> > > > > > now.
> > > > >
> > > > > Why do I just not need it?
> > > >
> > > > Let me flip it around, why do you need it?  There are no in-kernel users
> > > > of the return value anymore so what code requires this pointer now?
> > >
> > > There still are a few users of other members in the family, and some
> > > of them are meant to be removed without removing the full parent
> > > directory.  Having all debugfs_create_*() functions behave the same
> > > is a bonus.
> >
> > I agree, and we are almost there, all that is left is:
> >         debugfs_create_blob()
> >         debugfs_create_file()
> >         debugfs_create_file_unsafe()
> > for creating debugfs files.
> >
> > There is still:
> >         debugfs_create_dir()
> >         debugfs_create_symlink()
> >         debugfs_create_automount()
> > for non-file creations that do not return void.
> >
> > The majority of the debugfs_create_* functions now do not return
> > anything.
> >
> > > But if other people are fine with having to call
> > > debugfs_remove(debugfs_lookup(...)), well, let it be like that...
> >
> > It saves at least a static variable, so what's not to like?  :)
> 
> Which is more than offset by the cost of the new debugfs_lookup() call...

Not when people were keeping a dentry-per-entry in lots of structures.
That's from the heap, this dentry pointer sits on the stack, or if we
are lucky, only in a register :)

thanks,

greg k-h