On Mon, Apr 29, 2024 at 09:29:58AM -0700, Christoph Lameter (Ampere) wrote: > On Mon, 29 Apr 2024, Dan Carpenter wrote: > > > I've always thought freeing pointers that have not been allocated is > > sloppy so I like that kfree() doesn't allow error pointers. We always > > catch it before it reaches production and that teaches people better > > habbits. Personally, I like how free_netdev() only accepts valid > > pointers. > > kfree() already checks for NULL and ZERO pointers. We should add these > checks in *one* location. > > Maybe issue a WARN_ONCE() and simply treat it as a NULL pointer if an error > code is passed? Did you even read the initial patch? The point is that this new automatic destructor path can pass error pointers to the destructor for completely valid code. Warning would completely defeat the purpose of this exercise.
