Use OPENJDK_TARGET_CPU_ARCH instead of legacy value for hotspot ARCH

When configuring the hotspot build the build system sets up the ARCH variable to reflect the target cpu. Currently the value is initialized to OPENJDK_TARGET_CPU_LEGACY, which is the internal legacy cpu name. For example, on x86 64-bit this is amd64 on linux (but x86_64 on mac).

Depending on the platform ARCH is either honored or ignored. Solaris for example overrides the value of ARCH with the output from 'uname -p' (see hotspot/make/solaris/makefiles/defs.make). On BSD/OSX the value is also overidden, but with the output from 'uname -m' (see hotspot/make/bsd/makefiles/defs.make). Windows is.. different.

On Linux the ARCH variable is honored, but if it is *not* set it is initialized to the output from 'uname -m' (see hotspot/make/linux/makefiles/defs.make).

The goal in the new (JDK) build system is to have the "legacy" value gradually removed in favor of the other variables. We should move towards setting ARCH to OPENJDK_TARGET_CPU_ARCH instead of OPENJDK_TARGET_CPU_LEGACY and make sure the Hotspot build system picks it up correctly.