https://gcc.gnu.org/g:d9945609388be31b71baf8fada291de593d732e9
commit d9945609388be31b71baf8fada291de593d732e9 Author: Michael Meissner <[email protected]> Date: Thu Oct 1 18:24:41 2026 -0400 Update ChangeLog.* Diff: --- gcc/ChangeLog.float | 46 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) diff --git a/gcc/ChangeLog.float b/gcc/ChangeLog.float index f91c4bb06b0f..df88e329054f 100644 --- a/gcc/ChangeLog.float +++ b/gcc/ChangeLog.float @@ -1,5 +1,51 @@ ==================== Branch work265-float, patch #221 ==================== +Generate 16-bit floating point truncate functions + +Add support for automatically converting the soft-ftp extend functions for +16-bit. + +2026-10-01 Michael Meissner <[email protected]> + +libgcc/ + + * config/rs6000/convert-16bit-fp-trunc: New perl script. + * config/rs6000/t-float16: Update. + * config/rs6000/truncsfdf-bf2.c: Delete. + * config/rs6000/truncsfdf-hf2.c: Likewise. + +==================== Branch work265-float, patch #223 ==================== + +Generate 16-bit floating extend functions + +Add support for automatically converting the soft-ftp extend functions for +16-bit. Also rework the pevious conversion support. + +2026-10-01 Michael Meissner <[email protected]> + +libgcc/ + + * config/rs6000/convert-16bit-fp-extend: New perl script. + * config/rs6000/convert-16bit-fp-fix: Rewrite to use -t conversion + instead of -b/-h. + * config/rs6000/convert-16bit-fp-float: Likewise. + * config/rs6000/t-float16: Update. + * config/rs6000/extendbf-sfdf2.c: Delete. + * config/rs6000/extendhf-sfdf2.c: Likewise. + +==================== Branch work265-float, patch #221 ==================== + +Add back missing files. + +2026-10-01 Michael Meissner <[email protected]> + + * config/rs6000/ppc-fixbfbitint.c: Add file back. + * config/rs6000/ppc-fixhfbitint.c: Likewise. + * config/rs6000/ppc-floatbitintbf.c: Likewise. + * config/rs6000/ppc-floatbitinthf.c: Likewise. + +==================== Branch work265-float, patch #221 ==================== + Use perl to convert float functions for 16-bit floating point This patch removes the ppc-float* functions, and instead uses perl to generate
