Don't use Long1 for stencil operations.

src/Renderer/Surface.cpp

@@ -3051,6 +3051,7 @@ namespace sw
 		int height2 = (height + 1) & ~1;
 
 		// FIXME: Unpacking byte4 to short4 in the sampler currently involves reading 8 bytes,
+		// and stencil operations also read 8 bytes per four 8-bit stencil values,
 		// so we have to allocate 4 extra bytes to avoid buffer overruns.
 		return allocateZero(size(width2, height2, depth, format) + 4);
 	}

src/Shader/PixelRoutine.cpp

@@ -321,7 +321,7 @@ namespace sw
 			buffer += q * *Pointer<Int>(data + OFFSET(DrawData,stencilSliceB));
 		}
 
-		Byte8 value = As<Byte8>(Long1(*Pointer<UInt>(buffer)));
+		Byte8 value = *Pointer<Byte8>(buffer);
 		Byte8 valueCCW = value;
 
 		if(!state.noStencilMask)
@@ -763,7 +763,7 @@ namespace sw
 			buffer += q * *Pointer<Int>(data + OFFSET(DrawData,stencilSliceB));
 		}
 
-		Byte8 bufferValue = As<Byte8>(Long1(*Pointer<UInt>(buffer)));
+		Byte8 bufferValue = *Pointer<Byte8>(buffer);
 
 		Byte8 newValue;
 		stencilOperation(newValue, bufferValue, state.stencilPassOperation, state.stencilZFailOperation, state.stencilFailOperation, false, zMask, sMask);