Merge branch 'gallium-strict-aliasing'

4 files changed, 47 insertions, 35 deletions

diff --git a/src/mesa/main/imports.c b/src/mesa/main/imports.c
index c9e00cf752..f2ef84f35c 100644
--- a/src/mesa/main/imports.c
+++ b/src/mesa/main/imports.c
@@ -460,7 +460,7 @@ _mesa_inv_sqrtf(float n)
 #if 0 /* not used, see below -BP */
         float r3, x3, y3;
 #endif
-        union { float f; unsigned int i; } u;
+        fi_type u;
         unsigned int magic;
 
         /*
@@ -649,10 +649,10 @@ _mesa_bitcount(unsigned int n)
 GLhalfARB
 _mesa_float_to_half(float val)
 {
-   const int flt = *((int *) (void *) &val);
-   const int flt_m = flt & 0x7fffff;
-   const int flt_e = (flt >> 23) & 0xff;
-   const int flt_s = (flt >> 31) & 0x1;
+   const fi_type fi = {val};
+   const int flt_m = fi.i & 0x7fffff;
+   const int flt_e = (fi.i >> 23) & 0xff;
+   const int flt_s = (fi.i >> 31) & 0x1;
    int s, e, m = 0;
    GLhalfARB result;
    
@@ -739,7 +739,8 @@ _mesa_half_to_float(GLhalfARB val)
    const int m = val & 0x3ff;
    const int e = (val >> 10) & 0x1f;
    const int s = (val >> 15) & 0x1;
-   int flt_m, flt_e, flt_s, flt;
+   int flt_m, flt_e, flt_s;
+   fi_type fi;
    float result;
 
    /* sign bit */
@@ -774,8 +775,8 @@ _mesa_half_to_float(GLhalfARB val)
       flt_m = m << 13;
    }
 
-   flt = (flt_s << 31) | (flt_e << 23) | flt_m;
-   result = *((float *) (void *) &flt);
+   fi.i = (flt_s << 31) | (flt_e << 23) | flt_m;
+   result = fi.f;
    return result;
 }
 
diff --git a/src/mesa/main/macros.h b/src/mesa/main/macros.h
index 3d9a1aba98..f0ea463fb9 100644
--- a/src/mesa/main/macros.h
+++ b/src/mesa/main/macros.h
@@ -202,17 +202,12 @@ do {                                \
 #endif
 
 /**
- * Copy a 4-element float vector (avoid using FPU registers)
- * XXX Could use two 64-bit moves on 64-bit systems
+ * Copy a 4-element float vector
+ * memcpy seems to be most efficient
  */
 #define COPY_4FV( DST, SRC )                  \
 do {                                          \
-   const GLuint *_s = (const GLuint *) (SRC); \
-   GLuint *_d = (GLuint *) (DST);             \
-   _d[0] = _s[0];                             \
-   _d[1] = _s[1];                             \
-   _d[2] = _s[2];                             \
-   _d[3] = _s[3];                             \
+   _mesa_memcpy(DST, SRC, sizeof(GLfloat) * 4);       \
 } while (0)
 
 /** Copy \p SZ elements into a 4-element vector */
diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index 192d39aed1..038ce0b4f0 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -54,8 +54,18 @@
  * Set x to positive or negative infinity.
  */
 #if defined(USE_IEEE) || defined(_WIN32)
-#define SET_POS_INFINITY(x)  ( *((GLuint *) (void *)&x) = 0x7F800000 )
-#define SET_NEG_INFINITY(x)  ( *((GLuint *) (void *)&x) = 0xFF800000 )
+#define SET_POS_INFINITY(x)                  \
+   do {                                      \
+         fi_type fi;                         \
+         fi.i = 0x7F800000;                  \
+         x = fi.f;                           \
+   } while (0)
+#define SET_NEG_INFINITY(x)                  \
+   do {                                      \
+         fi_type fi;                         \
+         fi.i = 0xFF800000;                  \
+         x = fi.f;                           \
+   } while (0)
 #elif defined(VMS)
 #define SET_POS_INFINITY(x)  x = __MAXFLOAT
 #define SET_NEG_INFINITY(x)  x = -__MAXFLOAT
@@ -1635,11 +1645,12 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP2H:        /* unpack two 16-bit floats */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             GLhalfNV hx, hy;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            hx = rawBits[0] & 0xffff;
-            hy = rawBits[0] >> 16;
+            fi.f = a[0];
+            hx = fi.i & 0xffff;
+            hy = fi.i >> 16;
             result[0] = result[2] = _mesa_half_to_float(hx);
             result[1] = result[3] = _mesa_half_to_float(hy);
             store_vector4(inst, machine, result);
@@ -1648,11 +1659,12 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP2US:       /* unpack two GLushorts */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             GLushort usx, usy;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            usx = rawBits[0] & 0xffff;
-            usy = rawBits[0] >> 16;
+            fi.f = a[0];
+            usx = fi.i & 0xffff;
+            usy = fi.i >> 16;
             result[0] = result[2] = usx * (1.0f / 65535.0f);
             result[1] = result[3] = usy * (1.0f / 65535.0f);
             store_vector4(inst, machine, result);
@@ -1661,24 +1673,26 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP4B:        /* unpack four GLbytes */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            result[0] = (((rawBits[0] >> 0) & 0xff) - 128) / 127.0F;
-            result[1] = (((rawBits[0] >> 8) & 0xff) - 128) / 127.0F;
-            result[2] = (((rawBits[0] >> 16) & 0xff) - 128) / 127.0F;
-            result[3] = (((rawBits[0] >> 24) & 0xff) - 128) / 127.0F;
+            fi.f = a[0];
+            result[0] = (((fi.i >> 0) & 0xff) - 128) / 127.0F;
+            result[1] = (((fi.i >> 8) & 0xff) - 128) / 127.0F;
+            result[2] = (((fi.i >> 16) & 0xff) - 128) / 127.0F;
+            result[3] = (((fi.i >> 24) & 0xff) - 128) / 127.0F;
             store_vector4(inst, machine, result);
          }
          break;
       case OPCODE_UP4UB:       /* unpack four GLubytes */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            result[0] = ((rawBits[0] >> 0) & 0xff) / 255.0F;
-            result[1] = ((rawBits[0] >> 8) & 0xff) / 255.0F;
-            result[2] = ((rawBits[0] >> 16) & 0xff) / 255.0F;
-            result[3] = ((rawBits[0] >> 24) & 0xff) / 255.0F;
+            fi.f = a[0];
+            result[0] = ((fi.i >> 0) & 0xff) / 255.0F;
+            result[1] = ((fi.i >> 8) & 0xff) / 255.0F;
+            result[2] = ((fi.i >> 16) & 0xff) / 255.0F;
+            result[3] = ((fi.i >> 24) & 0xff) / 255.0F;
             store_vector4(inst, machine, result);
          }
          break;
diff --git a/src/mesa/state_tracker/st_atom_pixeltransfer.c b/src/mesa/state_tracker/st_atom_pixeltransfer.c
index 4b35f59cc2..6a5854e9ba 100644
--- a/src/mesa/state_tracker/st_atom_pixeltransfer.c
+++ b/src/mesa/state_tracker/st_atom_pixeltransfer.c
@@ -162,12 +162,14 @@ load_color_map_texture(GLcontext *ctx, struct pipe_texture *pt)
     */
    for (i = 0; i < texSize; i++) {
       for (j = 0; j < texSize; j++) {
+         union util_color uc;
          int k = (i * texSize + j);
          ubyte r = ctx->PixelMaps.RtoR.Map8[j * rSize / texSize];
          ubyte g = ctx->PixelMaps.GtoG.Map8[i * gSize / texSize];
          ubyte b = ctx->PixelMaps.BtoB.Map8[j * bSize / texSize];
          ubyte a = ctx->PixelMaps.AtoA.Map8[i * aSize / texSize];
-         util_pack_color_ub(r, g, b, a, pt->format, dest + k);
+         util_pack_color_ub(r, g, b, a, pt->format, &uc);
+         *(dest + k) = uc.ui;
       }
    }