Hi Alexander, On Fri, Jul 13, 2018 at 05:51:56PM +0200, Alexander Sverdlin wrote: > prom_putchar() is used centrally in early printk infrastructure therefore > at least MIPS should agree on the function return type. > > Signed-off-by: Alexander Sverdlin <alexander.sverdlin@xxxxxxxxx> > --- > arch/mips/ar7/prom.c | 3 +-- > arch/mips/boot/compressed/uart-prom.c | 3 +-- > arch/mips/cavium-octeon/setup.c | 3 +-- > arch/mips/fw/arc/arc_con.c | 1 + > arch/mips/include/asm/setup.h | 1 + > arch/mips/include/asm/sgialib.h | 1 - > arch/mips/include/asm/txx9/generic.h | 1 - > arch/mips/kernel/early_printk.c | 2 -- > arch/mips/paravirt/serial.c | 4 +--- > arch/mips/pic32/pic32mzda/early_console.c | 4 +--- > 10 files changed, 7 insertions(+), 16 deletions(-) I was wondering about whether asm/setup.h is the best place for the declaration of prom_putchar(), and whether asm/prom.h or asm/bootinfo.h (which declares a couple of other common prom_* functions) might be better suited, but ultimately I think asm/setup.h is fine given that it keeps the early printk declarations together. We could probably use some cleanup of our various headers with not-very-well-defined purpose at some point though... Anyway I've applied this to mips-next for 4.19, but with the addition of including asm/setup.h in all files which use or define prom_putchar() (and including linux/types.h to make sure we have the definition of bool which asm/setup.h uses). Thanks, Paul