From: José Fonseca <jfonseca@vmware.com>
Date: Thu, 20 Dec 2012 12:03:45 +0000 (+0000)
Subject: util/u_format: Round when converting depth values from float to z16_unorm.
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=74f07319533d39a509ddc3159a51cfe89cd47225;p=mesa.git

util/u_format: Round when converting depth values from float to z16_unorm.

This makes the z16_unorm -> float -> z16_unorm conversion lossless.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
---

diff --git a/src/gallium/auxiliary/util/u_format_zs.c b/src/gallium/auxiliary/util/u_format_zs.c
index 816ef14204c..ed45c52f9c9 100644
--- a/src/gallium/auxiliary/util/u_format_zs.c
+++ b/src/gallium/auxiliary/util/u_format_zs.c
@@ -72,7 +72,7 @@ static INLINE uint16_t
 z32_float_to_z16_unorm(float z)
 {
    const float scale = 0xffff;
-   return (uint16_t)(z * scale);
+   return (uint16_t)(z * scale + 0.5f);
 }
 
 static INLINE float