projects
/
gcc.git
/ commitdiff
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
| commitdiff |
tree
raw
|
patch
| inline |
side by side
(parent:
a74853c
)
Fix -msoft-float on Power.
author
Michael Meissner
<meissner@gcc.gnu.org>
Thu, 9 Mar 1995 17:51:37 +0000
(17:51 +0000)
committer
Michael Meissner
<meissner@gcc.gnu.org>
Thu, 9 Mar 1995 17:51:37 +0000
(17:51 +0000)
From-SVN: r9144
gcc/config/rs6000/rs6000.md
patch
|
blob
|
history
diff --git
a/gcc/config/rs6000/rs6000.md
b/gcc/config/rs6000/rs6000.md
index ae27254c4401d372c97ceaf50ad3e26768a56c82..170ae1cbbed7138c43080c1767e6521956ccd6ec 100644
(file)
--- a/
gcc/config/rs6000/rs6000.md
+++ b/
gcc/config/rs6000/rs6000.md
@@
-3895,7
+3895,7
@@
/* If operands[1] is a register, it may have double-precision data
in it, so truncate it to single precision. We need not do
this for POWERPC. */
- if (! TARGET_POWERPC && GET_CODE (operands[1]) == REG)
+ if (! TARGET_POWERPC &&
TARGET_HARD_FLOAT &&
GET_CODE (operands[1]) == REG)
{
rtx newreg = reload_in_progress ? operands[1] : gen_reg_rtx (SFmode);
emit_insn (gen_truncdfsf2 (newreg,