[Qemu-devel] [PATCH 6/6] softfloat: Fix missing inexact for floating-point add

Richard Henderson posted 6 patches 7 years, 2 months ago
[Qemu-devel] [PATCH 6/6] softfloat: Fix missing inexact for floating-point add
Posted by Richard Henderson 7 years, 2 months ago
For 0x1.0000000000003p+0 + 0x1.ffffffep+14 = 0x1.0001fffp+15
we dropped the sticky bit and so failed to raise inexact.

Reported-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 fpu/softfloat.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fpu/softfloat.c b/fpu/softfloat.c
index 8cd2400081..7d63cffdeb 100644
--- a/fpu/softfloat.c
+++ b/fpu/softfloat.c
@@ -701,7 +701,7 @@ static FloatParts addsub_floats(FloatParts a, FloatParts b, bool subtract,
             }
             a.frac += b.frac;
             if (a.frac & DECOMPOSED_OVERFLOW_BIT) {
-                a.frac >>= 1;
+                shift64RightJamming(a.frac, 1, &a.frac);
                 a.exp += 1;
             }
             return a;
-- 
2.17.1


Re: [Qemu-devel] [PATCH 6/6] softfloat: Fix missing inexact for floating-point add
Posted by Laurent Desnogues 7 years, 2 months ago
On Fri, Aug 10, 2018 at 9:31 PM, Richard Henderson
<richard.henderson@linaro.org> wrote:
> For 0x1.0000000000003p+0 + 0x1.ffffffep+14 = 0x1.0001fffp+15
> we dropped the sticky bit and so failed to raise inexact.
>
> Reported-by: Laurent Desnogues <laurent.desnogues@gmail.com>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Tested-by: Laurent Desnogues <laurent.desnogues@gmail.com>
Reviewed-by: Laurent Desnogues <laurent.desnogues@gmail.com>

Laurent

> ---
>  fpu/softfloat.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fpu/softfloat.c b/fpu/softfloat.c
> index 8cd2400081..7d63cffdeb 100644
> --- a/fpu/softfloat.c
> +++ b/fpu/softfloat.c
> @@ -701,7 +701,7 @@ static FloatParts addsub_floats(FloatParts a, FloatParts b, bool subtract,
>              }
>              a.frac += b.frac;
>              if (a.frac & DECOMPOSED_OVERFLOW_BIT) {
> -                a.frac >>= 1;
> +                shift64RightJamming(a.frac, 1, &a.frac);
>                  a.exp += 1;
>              }
>              return a;
> --
> 2.17.1
>