On Thu, May 10, 2018 at 12:03 PM, Khem Raj <raj.khem@xxxxxxxxx> wrote: > > > On 5/10/18 2:34 AM, Matt Redfearn wrote: >> >> Hi Khem, >> >> On 04/05/18 20:05, Khem Raj wrote: >>> >>> This warning is seen with gcc-8 >>> >>> error: 'sys_cachectl' alias between functions of incompatible types >>> 'long int(char *, int, int)' >>> and 'long int(long int, long int, long int)' >>> >>> Signed-off-by: Khem Raj <raj.khem@xxxxxxxxx> >>> Cc: Ralf Baechle <ralf@xxxxxxxxxxxxxx> >>> Cc: linux-mips@xxxxxxxxxxxxxx >>> --- >>> arch/mips/kernel/Makefile | 2 ++ >>> arch/mips/mm/Makefile | 1 + >>> 2 files changed, 3 insertions(+) >>> >>> diff --git a/arch/mips/kernel/Makefile b/arch/mips/kernel/Makefile >>> index f10e1e15e1c6..eb92e52eb3db 100644 >>> --- a/arch/mips/kernel/Makefile >>> +++ b/arch/mips/kernel/Makefile >>> @@ -2,6 +2,8 @@ >>> # >>> # Makefile for the Linux/MIPS kernel. >>> # >>> +CFLAGS_signal.o += $(call cc-disable-warning, attribute-alias) >>> +CFLAGS_syscall.o += $(call cc-disable-warning, attribute-alias) >> >> >> Rather than disabling the (potentially useful) warning for the whole >> compilation unit, the better fix, I believe, would be something like Arnd >> proposed https://patchwork.kernel.org/patch/10093317/ to disable it just for >> the definition of the syscall entry point . Whatever happened to that RFC >> Arnd? >> > > I tend to agree. I think, I have also tested a patch where I am manually > ignoring the warning via pragma around the function which essentially is > similar to what Arnd has proposed only that Arnd's patch is generic Sorry I never followed up on my initial patches. I still think we should do it that way, but will need some more time until I can revisit them, as I'm currently travelling. If someone else wants to pick up my patches and post a new version in the meantime, that would be greatly appreciated. Arnd
