https://gcc.gnu.org/g:1799fdeaa4f74ce7dd79f5544715219d9c1404a1
commit 1799fdeaa4f74ce7dd79f5544715219d9c1404a1 Author: Michael Meissner <[email protected]> Date: Wed Sep 16 18:06:00 2026 -0400 Update ChangeLog.* Diff: --- gcc/ChangeLog.float | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/gcc/ChangeLog.float b/gcc/ChangeLog.float index e59b814b9fd2..a1ce900ed454 100644 --- a/gcc/ChangeLog.float +++ b/gcc/ChangeLog.float @@ -1,3 +1,26 @@ +==================== Branch work263-float, patch #215 ==================== + +Re-enable power8 support; Add conversions to/from integers; Fix typo. + +2026-09-16 Michael Meissner <[email protected]> + +gcc/ + + * config/rs6000/float16.md (extendbfhf2): Correct insn spelling. + (fix_trunc<FP16:mode><GPR:mode>2): Add support for converting integers + to/from 16-bit floating point. + (fixuns_trunc<FP16:mode><GPR:mode>2): Likewise. + (float<GPR:mode><FP16:mode>2): Likewise. + (floatuns<GPR:mode><FP16:mode>2): Likewise. + * config/rs6000/rs6000.cc (rs6000_option_override_internal): Reallow + power8 to run 16-bit floating point. + +gcc/testsuite/ + + * gcc.target/powerpc/bf16-2.c: Add power8 test again. Only run test on + power10. Force power10 code to use power10 instructions. + (gcc.target/powerpc/float16-2.c): Likewise. + ==================== Branch work263-float, patch #214 ==================== Drop power8 bfloat16 test.
