i965: Improve the hacks for ARB_fp scalar^scalar POW on gen6.

This is still awful, but my ability to care about reworking the old
backend so we can just get a temporary value into a POW is awfully low
since the new backend does this all sensibly.

Fixes:
fp1-LIT test 1
fp1-LIT test 3 (case x < 0)
fp1-POW test (exponentiation)
fp-lit-mask

diff --git a/src/mesa/drivers/dri/i965/brw_wm_emit.c b/src/mesa/drivers/dri/i965/brw_wm_emit.c
index a8be10638c7eed19a50916b943374a263b30d5d1..be86e0e1287bb6ea3c50e77ff53528dbe4c86af1 100644 (file)
--- a/src/mesa/drivers/dri/i965/brw_wm_emit.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_emit.c
@@ -973,34 +973,23 @@ void emit_math2(struct brw_wm_compile *c,
       struct brw_reg temp_dst = dst[dst_chan];
 
       if (arg0[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
-        if (arg1[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
-           /* Both scalar arguments.  Do scalar calc. */
-           src0.hstride = BRW_HORIZONTAL_STRIDE_1;
-           src1.hstride = BRW_HORIZONTAL_STRIDE_1;
-           temp_dst.hstride = BRW_HORIZONTAL_STRIDE_1;
-           temp_dst.width = BRW_WIDTH_1;
-
-           if (arg0[0].subnr != 0) {
-              brw_MOV(p, temp_dst, src0);
-              src0 = temp_dst;
-
-              /* Ouch.  We've used the temp as a dst, and we still
-               * need a temp to store arg1 in, because src and dst
-               * offsets have to be equal.  Leaving this up to
-               * glsl2-965 to handle correctly.
-               */
-              assert(arg1[0].subnr == 0);
-           } else if (arg1[0].subnr != 0) {
-              brw_MOV(p, temp_dst, src1);
-              src1 = temp_dst;
-           }
-        } else {
-           brw_MOV(p, temp_dst, src0);
-           src0 = temp_dst;
-        }
-      } else if (arg1[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
-        brw_MOV(p, temp_dst, src1);
-        src1 = temp_dst;
+        brw_MOV(p, temp_dst, src0);
+        src0 = temp_dst;
+      }
+
+      if (arg1[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
+        /* This is a heinous hack to get a temporary register for use
+         * in case both arg0 and arg1 are constants.  Why you're
+         * doing exponentiation on constant values in the shader, we
+         * don't know.
+         *
+         * max_wm_grf is almost surely less than the maximum GRF, and
+         * gen6 doesn't care about the number of GRFs used in a
+         * shader like pre-gen6 did.
+         */
+        struct brw_reg temp = brw_vec8_grf(c->max_wm_grf, 0);
+        brw_MOV(p, temp, src1);
+        src1 = temp;
       }
 
       brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
@@ -1018,14 +1007,6 @@ void emit_math2(struct brw_wm_compile *c,
                   sechalf(src0),
                   sechalf(src1));
       }
-
-      /* Splat a scalar result into all the channels. */
-      if (arg0[0].hstride == BRW_HORIZONTAL_STRIDE_0 &&
-         arg1[0].hstride == BRW_HORIZONTAL_STRIDE_0) {
-        temp_dst.hstride = BRW_HORIZONTAL_STRIDE_0;
-        temp_dst.vstride = BRW_VERTICAL_STRIDE_0;
-        brw_MOV(p, dst[dst_chan], temp_dst);
-      }
    } else {
       GLuint saturate = ((mask & SATURATE) ?
                         BRW_MATH_SATURATE_SATURATE :