On Sun, Oct 18, 2020 at 11:50:10AM -0700, John Hubbard wrote: > On 10/18/20 9:18 AM, Linus Torvalds wrote: > > On Sat, Oct 17, 2020 at 4:14 PM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote: > > > > > > --- a/mm/gup_test.c~selftests-vm-use-a-common-gup_testh > > > +++ a/mm/gup_test.c > > > @@ -4,22 +4,7 @@ > > > +#include "../../../../mm/gup_test.h" > > > +++ a/mm/gup_test.h > > > > There is no way I'm applying a crazy patch like this. > > > > That can't be right. > > > > And even if it works, it *still* isn't right. > > > > Linus > > > > I feel the same way about the ridiculous ../..'s in the include path. But that > seemed to be the accepted way (I even have an email that I'm trying to find, in > which someone confirmed that...ah here, it is [1], and I'm adding Jason to CC) > for the odd situation of sharing a header file between kernel and user space, > for the kselftest world. > > If there's a different acceptable way, I'd be delighted to use that. See where it says: CFLAGS = -Wall -I ../../../../usr/include $(EXTRA_CFLAGS) Does adding an extra -I../../../../mm to that line not do the trick?
