On Wed, 2019-01-16 at 18:04 +0200, Kalle Valo wrote: > Lubomir Rintel <lkundrak@xxxxx> writes: > > > <asm/olpc.h> provides machine_is_olpc() stub for CONFIG_OLPC=n, > > compiler should just optimize the unneeded bits away. > > > > Signed-off-by: Lubomir Rintel <lkundrak@xxxxx> > > --- > > drivers/net/wireless/marvell/libertas/if_usb.c | 10 ---------- > > 1 file changed, 10 deletions(-) > > > > diff --git a/drivers/net/wireless/marvell/libertas/if_usb.c b/drivers/net/wireless/marvell/libertas/if_usb.c > > index 220dcdee8d2b..7b61c599948e 100644 > > --- a/drivers/net/wireless/marvell/libertas/if_usb.c > > +++ b/drivers/net/wireless/marvell/libertas/if_usb.c > > @@ -12,9 +12,7 @@ > > #include <linux/usb.h> > > #include <linux/olpc-ec.h> > > > > -#ifdef CONFIG_OLPC > > #include <asm/olpc.h> > > -#endif > > Using asm directory looked odd to me so I did some investigation: > > $ find -name olpc.h > ./arch/x86/include/asm/olpc.h > > Did you build this on something else than x86? For example, will the > compiler find olpc.h when building for arm? Uh, no. You're right, this indeed breaks on non-x86. Perhaps olpc-ec.h should provide something that could tell whether the EC is available that could be used instead of machine_is_olpc(), or the olpc.h moved outside arch/x86. For now, please just ignore the patch.. Lubo
