This is the mail archive of the
libc-ports@sources.redhat.com
mailing list for the libc-ports project.
Re: [PATCH v2 09/14] arm: Tidy architecture selection
On Fri, 1 Mar 2013, Richard Henderson wrote:
> +# elif defined(__ARM_ARCH_5T__) || defined(__ARM_ARCH_5TE__) \
> + defined(__ARM_ARCH_5TEJ__)
> +# define __ARM_ARCH 5
Or plain 5, or 5E (allowed by the architecture and -march=armv5 /
-march=armv5e, even if GCC doesn't know of any relevant -mcpu= processors
and such may not have existed).
> +# elif defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6ZK__) \
> + defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6T2__)
> +# define __ARM_ARCH 6
Or plain 6, or 6Z.
My reference for possible values is the union of the architecture columns
in GCC 4.7's arm-arches.def and arm-cores.def (given that for 4.8 and
later you have __ARM_ARCH predefined, so only the values known to 4.7 are
relevant here).
> +# elif defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) \
> + defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7EM__)
Or plain 7 (given that you're already covering values that aren't actually
compatible with building glibc).
OK fixed to handle all the other __ARM_ARCH_* values GCC might define, as
described above.
--
Joseph S. Myers
joseph@codesourcery.com