unimplemented: -mfloat-abi=hard and VFP?
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
unimplemented: -mfloat-abi=hard and VFP?
- To: arm-gnu@xxxxxxxxxxxxxxxx
- Subject: unimplemented: -mfloat-abi=hard and VFP?
- From: Ioi Lam <Ioi.Lam@xxxxxxx>
- Date: Sat, 06 Aug 2005 01:59:54 -0700
Hi,
Is VFP support not implemented in GCC yet? I am using the 2005Q1B
release (downloaded from ARM's web site)
/gcc/2005Q1B/bin/arm-none-linux-gnueabi-gcc -march=armv5te -mfpu=vfp
-mhard-float -O -o - -S vfp.c
vfp.c:1: sorry, unimplemented: -mfloat-abi=hard and VFP
.file "vfp.c"
.comm k,4,4
.ident "GCC: (GNU) 3.4.3 (release) (CodeSourcery ARM Q1B 2005)"
FPA seems to work fine, but it doesn't generate the VFP instructions (as
generated by ADS)
/gcc/2005Q1B/bin/arm-none-linux-gnueabi-gcc -march=armv5te -mfpu=fpa
-mhard-float -O -o - -S vfp.c
.file "vfp.c"
.text
.align 2
.global foo
.type foo, %function
foo:
@ args = 0, pretend = 0, frame = 0
@ frame_needed = 0, uses_anonymous_args = 0
@ link register save eliminated.
@ lr needed for prologue
str r0, [sp, #-4]!
ldfs f1, [sp], #4
adfs f0, f1, f1
fixz r0, f0
bx lr
.size foo, .-foo
The ADS output is:
||foo|| PROC
|L1.0|
FADDS s0,s0,s0
FSTMDBX sp!,{d8-d8}
FTOSIZS s16,s0
FMRS r0,s16
FLDMIAX sp!,{d8-d8}
MOV pc,lr
ENDP