From: Eric Anholt <eric@anholt.net>
Date: Sun, 22 Aug 2010 07:44:28 +0000 (-0700)
Subject: i965: Fix brw_math1 with scalar argument in gen6 FS.
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=250fccecc81e2ad8a63efaccae85793fbffd0a37;p=mesa.git

i965: Fix brw_math1 with scalar argument in gen6 FS.

The docs claim two conflicting things: One, that a scalar source is
supported.  Two, source hstride must be 1 and width must be exec size.
So splat a constant argument out into a full reg to operate on, since
violating the second set of constraints is clearly failing.

The alternative here might be to do a 1-wide exec on a constant
argument for math1.  It would probably save cycles too.  But I'll
leave that for the glsl2-965 branch.

Fixes glsl-algebraic-div-one-2.shader_test.
---

diff --git a/src/mesa/drivers/dri/i965/brw_wm_emit.c b/src/mesa/drivers/dri/i965/brw_wm_emit.c
index 5c07108eb82..05b440eaa83 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_emit.c
@@ -852,10 +852,22 @@ void emit_math1(struct brw_wm_compile *c,
 		const struct brw_reg *arg0)
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    int dst_chan = _mesa_ffs(mask & WRITEMASK_XYZW) - 1;
    GLuint saturate = ((mask & SATURATE) ?
 		      BRW_MATH_SATURATE_SATURATE :
 		      BRW_MATH_SATURATE_NONE);
+   struct brw_reg src;
+
+   if (intel->gen >= 6 && arg0[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
+      /* Gen6 math requires that source and dst horizontal stride be 1.
+       *
+       */
+      src = *dst;
+      brw_MOV(p, src, arg0[0]);
+   } else {
+      src = arg0[0];
+   }
 
    if (!(mask & WRITEMASK_XYZW))
       return; /* Do not emit dead code */
@@ -871,7 +883,7 @@ void emit_math1(struct brw_wm_compile *c,
 	    function,
 	    saturate,
 	    2,
-	    arg0[0],
+	    src,
 	    BRW_MATH_DATA_VECTOR,
 	    BRW_MATH_PRECISION_FULL);
 
@@ -882,7 +894,7 @@ void emit_math1(struct brw_wm_compile *c,
 	       function,
 	       saturate,
 	       3,
-	       sechalf(arg0[0]),
+	       sechalf(src),
 	       BRW_MATH_DATA_VECTOR,
 	       BRW_MATH_PRECISION_FULL);
    }