[Mesa-dev] [PATCH 1/5] u_format: Fix bit definition of UF10_MANTISSA_BITS.

[Ian Romanick]

On 10/26/2011 04:46 PM, Eric Anholt wrote:
> This is only used in the code for packing to INF, and resulted in an
> extra bit set that was set anyway, so it was harmless except for the
> confusion caused.

There Mesa fixes and the piglit test look good.

Reviewed-by: Ian Romanick <ian.d.romanick at intel.com>

> ---
>   src/gallium/auxiliary/util/u_format_r11g11b10f.h |    2 +-
>   1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/src/gallium/auxiliary/util/u_format_r11g11b10f.h b/src/gallium/auxiliary/util/u_format_r11g11b10f.h
> index 8e0572a..b1946f8 100644
> --- a/src/gallium/auxiliary/util/u_format_r11g11b10f.h
> +++ b/src/gallium/auxiliary/util/u_format_r11g11b10f.h
> @@ -37,7 +37,7 @@
>   #define UF10_EXPONENT_BIAS   15
>   #define UF10_EXPONENT_BITS   0x1F
>   #define UF10_EXPONENT_SHIFT  5
> -#define UF10_MANTISSA_BITS   0x3F
> +#define UF10_MANTISSA_BITS   0x1F
>   #define UF10_MANTISSA_SHIFT  (23 - UF10_EXPONENT_SHIFT)
>   #define UF10_MAX_EXPONENT    (UF10_EXPONENT_BITS<<  UF10_EXPONENT_SHIFT)
>