[patch][4/5] powerpc V2: Add the action of rounding for packing denormalized numbers

ebony.zhu at freescale.com ebony.zhu at freescale.com
Thu Feb 8 14:54:41 EST 2007


Add the action of rounding for packing denormalized
numbers, in order to fully comply with IEEE-754.

Signed-off-by:Ebony Zhu <ebony.zhu at freescale.com>
---
 arch/powerpc/math-emu/op-common.h |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/arch/powerpc/math-emu/op-common.h b/arch/powerpc/math-emu/op-common.h
index afb82b6..434e3dd 100644
--- a/arch/powerpc/math-emu/op-common.h
+++ b/arch/powerpc/math-emu/op-common.h
@@ -82,6 +82,7 @@ ({int __ret = 0;						\
 	if (X##_e <= _FP_WFRACBITS_##fs)			\
 	  {							\
 	    _FP_FRAC_SRS_##wc(X, X##_e, _FP_WFRACBITS_##fs);	\
+	    __ret |= _FP_ROUND(wc, X);				\
 	    _FP_FRAC_SLL_##wc(X, 1);				\
 	    if (_FP_FRAC_OVERP_##wc(fs, X))			\
 	      {							\
@@ -100,6 +101,7 @@ ({int __ret = 0;						\
 	    /* underflow to zero */				\
 	    X##_e = 0;						\
 	    _FP_FRAC_SET_##wc(X, _FP_ZEROFRAC_##wc);		\
+	    __ret |= _FP_ROUND(wc, X);				\
             __ret |= EFLAG_UNDERFLOW;				\
 	  }							\
       }								\
-- 
1.4.0




More information about the Linuxppc-dev mailing list