From: Ilia Mirkin <imirkin@alum.mit.edu>
Date: Tue, 29 Dec 2015 20:05:34 +0000 (-0500)
Subject: nv50/ir: float(s32 & 0xff) = float(u8), not s8
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=724134f68322087ef88bc590febd0011167ae367;p=mesa.git

nv50/ir: float(s32 & 0xff) = float(u8), not s8

Make sure to make conversion unsigned when we're ANDing the high bits
away. Fixes corruption in dolphin.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: "11.0 11.1" <mesa-stable@lists.freedesktop.org>
---

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
index 022626ccb8f..c2842c2186f 100644
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
@@ -1889,6 +1889,9 @@ AlgebraicOpt::handleCVT_EXTBF(Instruction *cvt)
          arg = shift->getSrc(0);
          offset = imm.reg.data.u32;
       }
+      // We just AND'd the high bits away, which means this is effectively an
+      // unsigned value.
+      cvt->sType = TYPE_U32;
    } else if (insn->op == OP_SHR &&
               insn->sType == cvt->sType &&
               insn->src(1).getImmediate(imm)) {