public inbox for gcc-cvs@sourceware.org help / color / mirror / Atom feed
From: Michael Meissner <meissner@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org Subject: [gcc(refs/users/meissner/heads/work164-bugs)] Revert all changes Date: Fri, 12 Apr 2024 07:25:12 +0000 (GMT) [thread overview] Message-ID: <20240412072512.8BED5385842C@sourceware.org> (raw) https://gcc.gnu.org/g:d97e784f9887aece6d857e5790ea931ea6e044ee commit d97e784f9887aece6d857e5790ea931ea6e044ee Author: Michael Meissner <meissner@linux.ibm.com> Date: Fri Apr 12 03:25:09 2024 -0400 Revert all changes Diff: --- gcc/config/rs6000/rs6000.md | 44 ++++---------------------------------------- 1 file changed, 4 insertions(+), 40 deletions(-) diff --git a/gcc/config/rs6000/rs6000.md b/gcc/config/rs6000/rs6000.md index 251b5c486b0..f1f120199f3 100644 --- a/gcc/config/rs6000/rs6000.md +++ b/gcc/config/rs6000/rs6000.md @@ -6041,10 +6041,6 @@ ; with a '#' template, and a define_split (with C code). The idea is ; to allow constant folding with the template of the define_insn, ; then to have the insns split later (between sched1 and final). -; -; If we have direct support for SImode in floating point registers, just -; convert the SImode value to DImode. If we are loading the value from memory, -; we will use the LFIWAX/LXSIWAX (define_expand "floatsidf2" [(parallel [(set (match_operand:DF 0 "gpc_reg_operand") @@ -6058,13 +6054,7 @@ { if (TARGET_LFIWAX && TARGET_FCFID) { - if (TARGET_POWERPC64 && TARGET_P8_VECTOR) - { - rtx di_tmp = convert_to_mode (DImode, operands[1], false); - emit_insn (gen_floatdidf2 (operands[0], di_tmp)); - } - else - emit_insn (gen_floatsidf2_lfiwax (operands[0], operands[1])); + emit_insn (gen_floatsidf2_lfiwax (operands[0], operands[1])); DONE; } else if (TARGET_FCFID) @@ -6120,10 +6110,6 @@ ;; If we don't have a direct conversion to single precision, don't enable this ;; conversion for 32-bit without fast math, because we don't have the insn to ;; generate the fixup swizzle to avoid double rounding problems. -; -; If we have direct support for SImode in floating point registers, just -; convert the SImode value to DImode. If we are loading the value from memory, -; we will use the LFIWAX/LXSIWAX (define_expand "floatunssisf2" [(set (match_operand:SF 0 "gpc_reg_operand") (unsigned_float:SF (match_operand:SI 1 "nonimmediate_operand")))] @@ -6134,13 +6120,7 @@ { if (TARGET_LFIWZX && TARGET_FCFIDUS) { - if (TARGET_POWERPC64 && TARGET_P8_VECTOR) - { - rtx di_tmp = convert_to_mode (DImode, operands[1], true); - emit_insn (gen_floatdisf2 (operands[0], di_tmp)); - } - else - emit_insn (gen_floatunssisf2_lfiwzx (operands[0], operands[1])); + emit_insn (gen_floatunssisf2_lfiwzx (operands[0], operands[1])); DONE; } else @@ -6165,13 +6145,7 @@ { if (TARGET_LFIWZX && TARGET_FCFID) { - if (TARGET_POWERPC64 && TARGET_P8_VECTOR) - { - rtx di_tmp = convert_to_mode (DImode, operands[1], false); - emit_insn (gen_floatunsdidf2 (operands[0], di_tmp)); - } - else - emit_insn (gen_floatunssidf2_lfiwzx (operands[0], operands[1])); + emit_insn (gen_floatunssidf2_lfiwzx (operands[0], operands[1])); DONE; } else if (TARGET_FCFID) @@ -6931,10 +6905,6 @@ ;; If we don't have a direct conversion to single precision, don't enable this ;; conversion for 32-bit without fast math, because we don't have the insn to ;; generate the fixup swizzle to avoid double rounding problems. -; -; If we have direct support for SImode in floating point registers, just -; convert the SImode value to DImode. If we are loading the value from memory, -; we will use the LFIWAX/LXSIWAX (define_expand "floatsisf2" [(set (match_operand:SF 0 "gpc_reg_operand") (float:SF (match_operand:SI 1 "nonimmediate_operand")))] @@ -6945,13 +6915,7 @@ { if (TARGET_FCFIDS && TARGET_LFIWAX) { - if (TARGET_POWERPC64 && TARGET_P8_VECTOR) - { - rtx di_tmp = convert_to_mode (DImode, operands[1], false); - emit_insn (gen_floatdisf2 (operands[0], di_tmp)); - } - else - emit_insn (gen_floatsisf2_lfiwax (operands[0], operands[1])); + emit_insn (gen_floatsisf2_lfiwax (operands[0], operands[1])); DONE; } else if (TARGET_FCFID && TARGET_LFIWAX)
next reply other threads:[~2024-04-12 7:25 UTC|newest] Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top 2024-04-12 7:25 Michael Meissner [this message] -- strict thread matches above, loose matches on Subject: below -- 2024-04-12 16:26 Michael Meissner 2024-04-09 21:37 Michael Meissner
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20240412072512.8BED5385842C@sourceware.org \ --to=meissner@gcc.gnu.org \ --cc=gcc-cvs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).