Option -mfloat-gprs=yes works fine with basic operations on floats like +, -, /
*.
The following simple example produces an internal compiler error:
# 1 "test.c"
# 1 "<built-in>"
# 1 "<command line>"
# 1 "test.c"
int compare(float x, float y)
{
return (x < y);
}
The error message is
test.c: In function `compare':
test.c:5: error: unrecognizable insn:
(insn 14 13 15 0 (set (reg:CCFP 122)
(compare:CCFP (reg:SF 120 [ x ])
(reg:SF 121 [ y ]))) -1 (nil)
(expr_list:REG_DEAD (reg:SF 120 [ x ])
(expr_list:REG_DEAD (reg:SF 121 [ y ])
(nil))))
test.c:5: internal compiler error: in extract_insn, at recog.c:2083
Please submit a full bug report,
with preprocessed source if appropriate.
The compile command is
powerpc-eabispe-gcc -bpowerpc-eabispe -mcpu=505 -mno-sdata -mfloat-gprs=yes
-Wall -Wstrict-prototypes -Wall -save-temps -c -o o/test.o test.c
The gcc version output is
Reading specs from /opt/CSE/lib/gcc/powerpc-eabispe/3.4.6/specs
Configured with: ../gcc-3.4.6/configure --disable-nls --prefix=/opt/CSE
--target=powerpc-eabispe
Thread model: single
gcc version 3.4.6
--
Summary: internal compiler error, if float-comparison is done
with option -mfloat-gprs=yes
Product: gcc
Version: 3.4.6
Status: UNCONFIRMED
Severity: blocker
Priority: P3
Component: c
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: poellmann at nm dot hsd dot utc dot com
GCC build triplet: ../gcc-3.4.6/configure --disable-nls --prefix=/opt/CSE -
-target=
GCC host triplet: Cygwin on Win2000
GCC target triplet: powerpc-eabispe
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29815
