public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc(refs/users/meissner/heads/work095)] Allow IBM 128-bit to do offset addressing.
@ 2022-07-23  0:11 Michael Meissner
  0 siblings, 0 replies; only message in thread
From: Michael Meissner @ 2022-07-23  0:11 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:579acbc3080ad5d10cf8ff09f1b8cfb3aa4fe031

commit 579acbc3080ad5d10cf8ff09f1b8cfb3aa4fe031
Author: Michael Meissner <meissner@linux.ibm.com>
Date:   Fri Jul 22 20:10:13 2022 -0400

    Allow IBM 128-bit  to do offset addressing.
    
    I noticed that we didn't enable offset addressing for IBM 128-bit floating
    point.
    
    2022-07-22   Michael Meissner  <meissner@linux.ibm.com>
    
    gcc/
    
            * config/rs6000/rs6000.cc (rs6000_setup_reg_addr_masks): Allow
            IFmode and possibly TFmode to use offset addressing.
            (reg_offset_addressing_ok_p): IBM 128-bit floating point modes can do
            offset addressing.

Diff:
---
 gcc/config/rs6000/rs6000.cc | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)

diff --git a/gcc/config/rs6000/rs6000.cc b/gcc/config/rs6000/rs6000.cc
index 667f83b1dfd..ae61505123d 100644
--- a/gcc/config/rs6000/rs6000.cc
+++ b/gcc/config/rs6000/rs6000.cc
@@ -2697,6 +2697,11 @@ rs6000_setup_reg_addr_masks (void)
 			  || (rc == RELOAD_REG_VMX && TARGET_P9_VECTOR)))))
 	    addr_mask |= RELOAD_REG_OFFSET;
 
+	  /* IBM 128-bit can do REG+OFFSET addressing.  */
+	  else if ((addr_mask != 0) && !indexed_only_p
+		   && FLOAT128_IBM_P (m))
+	    addr_mask |= RELOAD_REG_OFFSET;
+
 	  /* VSX registers can do REG+OFFSET addresssing if ISA 3.0
 	     instructions are enabled.  The offset for 128-bit VSX registers is
 	     only 12-bits.  While GPRs can handle the full offset range, VSX
@@ -8511,6 +8516,16 @@ reg_offset_addressing_ok_p (machine_mode mode)
 {
   switch (mode)
     {
+    case E_IFmode:
+      return true;
+
+    case E_TFmode:
+      if (FLOAT128_IBM_P (TFmode))
+	return true;
+
+      /* If TFmode is IEEE 128-bit, treat it like a vector.  */
+      /* fall through */
+
     case E_V16QImode:
     case E_V8HImode:
     case E_V4SFmode:
@@ -8519,7 +8534,6 @@ reg_offset_addressing_ok_p (machine_mode mode)
     case E_V2DImode:
     case E_V1TImode:
     case E_TImode:
-    case E_TFmode:
     case E_KFmode:
       /* AltiVec/VSX vector modes.  Only reg+reg addressing was valid until the
 	 ISA 3.0 vector d-form addressing mode was added.  While TImode is not


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-07-23  0:11 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-23  0:11 [gcc(refs/users/meissner/heads/work095)] Allow IBM 128-bit to do offset addressing Michael Meissner

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).