Hi, This patch set builds on the AArch64 support for _Float16 that will be added once the patch set starting at https://gcc.gnu.org/ml/gcc-patches/2016-09/msg02383.html is accepted.
Unlike the AArch64 support, we've not got too much to do, outside of setting up a single step conversions between HFmode and DFmode. Patches 1-3 (co-authored with Matthew Wahab) add this support, by refactoring __gnu_f2h_internal, then expanding it to support 16-bit to 64-bit conversions, then removing TARGET_CONVERT_TO_TYPE and wiring up truncdfhf2 and extendhfdf2 for ARM. Patch 4 adds the requisite wiring to enable the _Float16 type on ARM. I've bootstrapped the patch set on ARMv8-A hardware, and cross-tested it with appropriate multilibs. OK once the AArch64 support is in? Thanks, James