PATCH: PR target/15383: __float128 function name is used for long double

[H. J. Lu]

See

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15383

The problem is a __float128 function name is used for long double for
backward compatibility. It was OK until gcc 3.4.0 supports __float128.
This patch removes it. Should I keep the old version for backward
compatibility? It will only work on Linux.


H.J.
-----
2004-05-11  H.J. Lu  <hongjiu.lu@intel.com>

	* config/ia64/lib1funcs.asm (__divtf3): Removed.

--- gcc/config/ia64/lib1funcs.asm.float128	2003-10-27 11:17:23.000000000 -0800
+++ gcc/config/ia64/lib1funcs.asm	2004-05-11 12:31:00.000000000 -0700
@@ -5,16 +5,12 @@
 // alternative.
 //
 // farg0 holds the dividend.  farg1 holds the divisor.
-//
-// __divtf3 is an alternate symbol name for backward compatibility.
 
 	.text
 	.align 16
 	.global __divxf3
-	.global __divtf3
 	.proc __divxf3
 __divxf3:
-__divtf3:
 	cmp.eq p7, p0 = r0, r0
 	frcpa.s0 f10, p6 = farg0, farg1
 	;;