On Sun, Jul 04, 2021 at 03:44:17PM -0700, Linus Torvalds wrote: > On Sun, Jul 4, 2021 at 2:47 PM Guenter Roeck <linux@xxxxxxxxxxxx> wrote: > > > > How about the following ? > > > > WARN_ON_ONCE(IS_ENABLED(CONFIG_MMU) && uaccess_kernel()); > > Nope, that doesn't work either, because there are no-MMU setups that > don't make the same mistake no-mmu arm and m68k do. > > Example: xtensa. But afaik also generic-asm/uaccess.h unless the > architecture overrides something. > > So this literally seems like just an arm/m68k bug. We could slip: #ifndef uaccess_user #define uaccess_user() !uaccess_kernel() #endif into asm-generic, switch the test over and then make it arm/m68k's problem to define uaccess_user() to true?
