This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Possible ARM binutils/gcc errors...
- To: Philip Blundell <philb at gnu dot org>
- Subject: Re: Possible ARM binutils/gcc errors...
- From: Russell King - ARM Linux <linux at arm dot linux dot org dot uk>
- Date: Thu, 2 Aug 2001 20:42:07 +0100
- Cc: sjhill at cotw dot com, linux-arm at lists dot arm dot linux dot org dot uk, binutils at sources dot redhat dot com
- References: <3B61AD6C.B2E311CF@cotw.com> <sjhill@cotw.com> <E15QDyu-0000IZ-00@kings-cross.london.uk.eu.org>
On Fri, Jul 27, 2001 at 09:13:04PM +0100, Philip Blundell wrote:
> The kernel isn't using the right flags for assembler files. Try patching
> arch/arm/Makefile like this.
>
> @@ -39,7 +38,7 @@ tune-$(CONFIG_CPU_SA110) :=-mtune=strong
> tune-$(CONFIG_CPU_SA1100) :=-mtune=strongarm1100
>
> CFLAGS += -mapcs-32 $(apcs-y) $(arch-y) $(tune-y) -mshort-load-bytes -msoft-float
> -AFLAGS += -mapcs-32 $(apcs-y) $(arch-y) -mno-fpu
> +AFLAGS += -mapcs-32 $(apcs-y) $(arch-y) -mno-fpu -msoft-float
For the record, I have no intention of accepting the above patch at
present; its another build-breaking change in the ARM toolchain.
There are two options:
1. Don't use the new toolchains to build kernels [recommended]
2. Come up with a decent test to see if -msoft-float works on the
toolchain.
Including it appears to prevent anything other than GCC 2.96 and up
from building kernels, and I don't regard these later toolchains to be
sufficiently well proven and mature to be worth the effort at the moment.