[PATCH, i386]: Committed: define WIDEST_HARDWARE_FP_SIZE

Uros Bizjak ubizjak@gmail.com
Mon Dec 17 01:12:00 GMT 2007


Hello!

As mentioned by Joseph in PR 34473, x86 target should define 
WIDEST_HARDWARE_FP_SIZE, otherwise some of int-float conversion 
functions from libgcc2 start using software TFmode instead of their 
fallback implementations.

Patch was bootstrapped and regression tested on x86_64 {,-m32}. Patch is 
commited to SVN.

2007-12-17  Uros Bizjak  <ubizjak@gmail.com>

        * config/i386/i386.h (WIDEST_HARDWARE_FP_SIZE): Define.

Uros.

Index: i386.h
===================================================================
--- i386.h      (revision 130989)
+++ i386.h      (working copy)
@@ -749,6 +749,8 @@
 
 #define LONG_DOUBLE_TYPE_SIZE 80
 
+#define WIDEST_HARDWARE_FP_SIZE 80
+
 /* Set the value of FLT_EVAL_METHOD in float.h.  When using only the
    FPU, assume that the fpcw is set to extended precision; when using
    only SSE, rounding is correct; when using both SSE and the FPU,



More information about the Gcc-patches mailing list