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/dmf001)] Update ChangeLog.meissner.
Date: Mon, 10 Oct 2022 21:52:07 +0000 (GMT)	[thread overview]
Message-ID: <20221010215208.35AED3858289@sourceware.org> (raw)

https://gcc.gnu.org/g:e79b6684d4f2074fcfcb72d0010bb4c0789b77f8

commit e79b6684d4f2074fcfcb72d0010bb4c0789b77f8
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Mon Oct 10 17:51:42 2022 -0400

    Update ChangeLog.meissner.
    
    2022-10-10   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * ChangeLog.meissner: Update.

Diff:
---
 gcc/ChangeLog.meissner | 107 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 107 insertions(+)

diff --git a/gcc/ChangeLog.meissner b/gcc/ChangeLog.meissner
index a0706e92964..dc19dd754a6 100644
--- a/gcc/ChangeLog.meissner
+++ b/gcc/ChangeLog.meissner
@@ -1,3 +1,110 @@
+==================== dmf001, patch #3
+
+Add initial DMR register support for XOmode.
+
+2022-10-10   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/mma.md (movxo_p10): Rename from movxo, and limit its use
+	to non-DMF systems.
+	(movxo_dmf): New insn.
+	(mma_<acc>): NOP prime/de-prime usage when -mdmf is used.
+	* config/rs6000/rs6000.cc (rs6000_hard_regno_mode_ok_uncached): Allow
+	XOmode in DMR registers or Altivec registers if -mdmf.
+	(rs6000_setup_reg_addr_masks): Add support to allow TDOmode in DMF
+	registers, and also in Altivec registers if -mdmf.  TDOmode registers
+	can do offset addressing but not indexed addressing.
+	(rs6000_init_hard_regno_mode_ok): Add support for TDOmode.
+	(reg_offset_addressing_ok_p): Likewise.
+	(rs6000_emit_move): Warn about using SUBREGs with TDOmode.
+	(rs6000_preferred_reload_class): DMF registers can't be loaded or
+	stored.  With -mdmf, XOmode can go in Altivec registers.  TDOmode needs
+	to use VSX registers for load/store.
+	(rs6000_compute_pressure_classes): Add DMF registers to the pressure
+	classses.
+	(rs6000_split_multireg_move): If -mdmf, we don't need to prime or
+	de-prime the acculators.  Add initial support for TDOmode.
+	(rs6000_invalid_conversion): Warn about converting __dmr types.
+	* config/rs6000/rs6000.h (DMF_REG_CLASS_P): New macro.
+
+==================== dmf001, patch #2
+
+Add initial DMR register support.
+
+2022-10-07   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/constraints.md (wD): New constraint.
+	* config/rs6000/predicates.md (dmf_operand): New predicate.
+	* config/rs6000/rs6000-modes.def (TDOmode): New opaque mode.
+	* config/rs6000/rs6000.cc (rs6000_reg_type): Add support for DMF
+	registers.
+	(rs6000_reload_reg_type): Likewise.
+	(LAST_RELOAD_REG_CLASS): Likewise.
+	(reload_reg_map): Likewise.
+	(addr_mask_type): Grow to 16 bits.
+	(RELOAD_REG_NO_MEMORY): New RELOAD_REG macro.
+	(rs6000_reg_names): Add DMR registers.
+	(alt_reg_names): Likewise.
+	(rs6000_hard_regno_nregs_internal): Add support for DMF registers.
+	(rs6000_hard_regno_mode_ok_uncached): Likewise.
+	(rs6000_modes_tieable_p): Likewise.
+	(rs6000_debug_addr_mask): Likewise.
+	(rs6000_debug_reg_global): Likewise.
+	(rs6000_setup_reg_addr_masks): Likewise.
+	(rs6000_init_hard_regno_mode_ok): Likewise.
+	* config/rs6000/rs6000.h (UNITS_PER_DMF_WORD): New macro.
+	(FIRST_PSEUDO_REGISTER): Add support for DMF registers.
+	(FIXED_REGISTERS): Likewise.
+	(CALL_REALLY_USED_REGISTERS): Likewise.
+	(REG_ALLOC_ORDER): Likewise.
+	(DMF_REGNO_P): New macro.
+	(enum reg_class): Likewise.
+	(REG_CLASS_NAMES): Likewise.
+	(REG_CLASS_CONTENTS): Likewise.
+	(REGISTER_NAMES): Likewise.
+	(ADDITIONAL_REGISTER_NAMES): Likewise.
+	* config/rs6000/rs6000.md (FIRST_DMF_REGNO): New constant.
+	(LAST_DMF_REGNO): New constant.
+	(isa attribute): Add dmf and mma_fpr attributes.
+	(enabled attribute): Add support for dmf and mma_fpr attributes.
+
+==================== dmf001, patch #3
+
+Add -mcpu=future/-mdmf instrastructure.
+
+2022-10-07   Michael Meissner  <meissner@linux.ibm.com>
+
+gcc/
+
+	* config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Define
+	__DMF__ if dmf support is enabled.
+	* config/rs6000/rs6000-cpus.def (ISA_FUTURE_MASKS): New macro.
+	(POWERPC_MASKS): Add -mblock-ops-vector-pair and -mdmf.
+	(future cpu): Add -mcpu=future and -mtune=future support.
+	* config/rs6000/rs6000-opts.h (processor_type): Add PROCESSOR_FUTURE.
+	* config/rs6000/rs6000-tables.op: Regenerate.
+	* config/rs6000/rs6000.cc (rs6000_option_override_internal): Add
+	-mcpu=future and -mtune=future support, but for now make them mostly
+	equivalent to power10.  Add -mdmf checking.
+	(rs6000_machine_from_flags): Likewise.
+	(rs6000_reassociation_width): Likewise.
+	(rs6000_adjust_cost): Likewise.
+	(rs6000_issue_rate): Likewise.
+	(rs6000_sched_reorder): Likewise.
+	(rs6000_sched_reorder2): Likewise.
+	(rs6000_register_move_cost): Likewise.
+	(rs6000_opt_masks): Add -mdmf.
+	* config/rs6000/rs6000.h (ASM_CPU_SPEC): Add -mcpu=future support.
+	* config/rs6000/rs6000.md (cpu attribute): Add future.
+	(isa attribute): Add dmf.
+	(enabled attribute): Likewise.
+	* config/rs6000/rs6000.opt (-mdmf): New option.
+
+==================== dmf001, base line
+
 2022-10-06   Michael Meissner  <meissner@linux.ibm.com>
 
 	Clone branch

             reply	other threads:[~2022-10-10 21:52 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-10 21:52 Michael Meissner [this message]
2022-10-10 21:52 Michael Meissner
2022-10-11 16:02 Michael Meissner
2022-10-11 23:19 Michael Meissner
2022-10-12  2:33 Michael Meissner
2022-10-13 17:59 Michael Meissner
2022-10-14 17:36 Michael Meissner
2022-10-15  0:06 Michael Meissner
2022-10-15  0:12 Michael Meissner
2022-10-17 23:18 Michael Meissner
2022-10-18 22:26 Michael Meissner
2022-10-18 22:36 Michael Meissner
2022-10-19  3:46 Michael Meissner
2022-10-19  3:53 Michael Meissner
2022-10-19  4:16 Michael Meissner
2022-10-20  4:11 Michael Meissner
2022-10-20  6:06 Michael Meissner
2022-10-27  3:37 Michael Meissner
2022-10-27  5:17 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=20221010215208.35AED3858289@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: link
Be 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).