Move span stuff from swrast.h to s_context.h (it's private).

Implemented remaining fragment program instructions.
Initial changes to implement fragment program texture sampling.

1 files changed, 154 insertions, 15 deletions

diff --git a/src/mesa/swrast/s_context.h b/src/mesa/swrast/s_context.h
index 3fd7f3c3a94..2be1545c8f3 100644
--- a/src/mesa/swrast/s_context.h
+++ b/src/mesa/swrast/s_context.h
@@ -1,10 +1,10 @@
-/* $Id: s_context.h,v 1.23 2003/01/14 04:55:46 brianp Exp $ */
+/* $Id: s_context.h,v 1.24 2003/02/23 04:10:54 brianp Exp $ */
 
 /*
  * Mesa 3-D graphics library
  * Version:  5.1
  *
- * Copyright (C) 1999-2002  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2003  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,14 +22,12 @@
  * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors:
- *    Keith Whitwell <[email protected]>
  */
 
+
 /**
  * \file swrast/s_context.h
- * \brief fill in description
+ * \brief Software rasterization context and private types.
  * \author Keith Whitwell <[email protected]>
  */
 
@@ -39,19 +37,160 @@
 #include "mtypes.h"
 #include "swrast.h"
 
-/*
- * For texture sampling:
+
+/**
+ * \struct sw_span
+ * \brief Contains data for either a horizontal line or a set of
+ * pixels that are passed through a pipeline of functions before being
+ * drawn.
+ *
+ * The sw_span structure describes the colors, Z, fogcoord, texcoords,
+ * etc for either a horizontal run or a set of independent pixels.  We
+ * can either specify a base/step to indicate interpolated values, or
+ * fill in arrays of values.  The interpMask and arrayMask bitfields
+ * indicate which are active.
+ *
+ * With this structure it's easy to hand-off span rasterization to
+ * subroutines instead of doing it all inline in the triangle functions
+ * like we used to do.
+ * It also cleans up the local variable namespace a great deal.
+ *
+ * It would be interesting to experiment with multiprocessor rasterization
+ * with this structure.  The triangle rasterizer could simply emit a
+ * stream of these structures which would be consumed by one or more
+ * span-processing threads which could run in parallel.
  */
-typedef void (*TextureSampleFunc)( GLcontext *ctx, GLuint texUnit,
-				   const struct gl_texture_object *tObj,
-                                   GLuint n, GLfloat texcoords[][4],
-                                   const GLfloat lambda[], GLchan rgba[][4] );
 
 
+/**
+ * \defgroup SpanFlags SPAN_XXX-flags
+ * Bitmasks to indicate which span_arrays need to be computed
+ * (sw_span::interpMask) or have already been filled
+ * (sw_span::arrayMask)
+ */
+/*@{*/
+#define SPAN_RGBA         0x001
+#define SPAN_SPEC         0x002
+#define SPAN_INDEX        0x004
+#define SPAN_Z            0x008
+#define SPAN_FOG          0x010
+#define SPAN_TEXTURE      0x020
+#define SPAN_INT_TEXTURE  0x040
+#define SPAN_LAMBDA       0x080
+#define SPAN_COVERAGE     0x100
+#define SPAN_FLAT         0x200  /**< flat shading? */
+/** sw_span::arrayMask only - for span_arrays::x, span_arrays::y */
+#define SPAN_XY           0x400
+#define SPAN_MASK         0x800  /**< sw_span::arrayMask only */
+/*@}*/
+
 
-/*
- * Blending function
+/**
+ * \struct span_arrays 
+ * \brief Arrays of fragment values.
+ *
+ * These will either be computed from the x/xStep values above or
+ * filled in by glDraw/CopyPixels, etc.
+ * These arrays are separated out of sw_span to conserve memory.
  */
+struct span_arrays {
+   GLchan  rgb[MAX_WIDTH][3];
+   GLchan  rgba[MAX_WIDTH][4];
+   GLuint  index[MAX_WIDTH];
+   GLchan  spec[MAX_WIDTH][4]; /* specular color */
+   GLint   x[MAX_WIDTH];  /**< X/Y used for point/line rendering only */
+   GLint   y[MAX_WIDTH];  /**< X/Y used for point/line rendering only */
+   GLdepth z[MAX_WIDTH];
+   GLfloat fog[MAX_WIDTH];
+   GLfloat texcoords[MAX_TEXTURE_COORD_UNITS][MAX_WIDTH][4];
+   GLfloat lambda[MAX_TEXTURE_COORD_UNITS][MAX_WIDTH];
+   GLfloat coverage[MAX_WIDTH];
+
+   /** This mask indicates if fragment is alive or culled */
+   GLubyte mask[MAX_WIDTH];
+};
+
+
+struct sw_span {
+   GLint x, y;
+
+   /** Only need to process pixels between start <= i < end */
+   /** At this time, start is always zero. */
+   GLuint start, end;
+
+   /** This flag indicates that mask[] array is effectively filled with ones */
+   GLboolean writeAll;
+
+   /** either GL_POLYGON, GL_LINE, GL_POLYGON, GL_BITMAP */
+   GLenum primitive;
+
+   /** 0 = front-facing span, 1 = back-facing span (for two-sided stencil) */
+   GLuint facing;
+
+   /**
+    * This bitmask (of  \link SpanFlags SPAN_* flags\endlink) indicates
+    * which of the x/xStep variables are relevant.
+    */
+   GLuint interpMask;
+
+#if CHAN_TYPE == GL_FLOAT
+   GLfloat red, redStep;
+   GLfloat green, greenStep;
+   GLfloat blue, blueStep;
+   GLfloat alpha, alphaStep;
+   GLfloat specRed, specRedStep;
+   GLfloat specGreen, specGreenStep;
+   GLfloat specBlue, specBlueStep;
+#else /* CHAN_TYPE == GL_UNSIGNED_BYTE or GL_UNSIGNED SHORT */
+   GLfixed red, redStep;
+   GLfixed green, greenStep;
+   GLfixed blue, blueStep;
+   GLfixed alpha, alphaStep;
+   GLfixed specRed, specRedStep;
+   GLfixed specGreen, specGreenStep;
+   GLfixed specBlue, specBlueStep;
+#endif
+   GLfixed index, indexStep;
+   GLfixed z, zStep;
+   GLfloat fog, fogStep;
+   GLfloat tex[MAX_TEXTURE_COORD_UNITS][4];  /* s, t, r, q */
+   GLfloat texStepX[MAX_TEXTURE_COORD_UNITS][4];
+   GLfloat texStepY[MAX_TEXTURE_COORD_UNITS][4];
+   GLfixed intTex[2], intTexStep[2];  /* s, t only */
+
+   /**
+    * This bitmask (of \link SpanFlags SPAN_* flags\endlink) indicates
+    * which of the fragment arrays in the span_arrays struct are relevant.
+    */
+   GLuint arrayMask;
+
+   /**
+    * We store the arrays of fragment values in a separate struct so
+    * that we can allocate sw_span structs on the stack without using
+    * a lot of memory.  The span_arrays struct is about 400KB while the
+    * sw_span struct is only about 512 bytes.
+    */
+   struct span_arrays *array;
+};
+
+
+#define INIT_SPAN(S, PRIMITIVE, END, INTERP_MASK, ARRAY_MASK)	\
+do {								\
+   (S).primitive = (PRIMITIVE);					\
+   (S).interpMask = (INTERP_MASK);				\
+   (S).arrayMask = (ARRAY_MASK);				\
+   (S).start = 0;						\
+   (S).end = (END);						\
+   (S).facing = 0;						\
+   (S).array = SWRAST_CONTEXT(ctx)->SpanArrays;			\
+} while (0)
+
+
+typedef void (*texture_sample_func)(GLcontext *ctx, GLuint texUnit,
+                                    const struct gl_texture_object *tObj,
+                                    GLuint n, const GLfloat texcoords[][4],
+                                    const GLfloat lambda[], GLchan rgba[][4]);
+
 #ifdef USE_MMX_ASM
 typedef void (_ASMAPIP blend_func)( GLcontext *ctx, GLuint n,
                                     const GLubyte mask[],
@@ -188,7 +327,7 @@ typedef struct
    /** Internal hooks, kept uptodate by the same mechanism as above.
     */
    blend_func BlendFunc;
-   TextureSampleFunc TextureSample[MAX_TEXTURE_IMAGE_UNITS];
+   texture_sample_func TextureSample[MAX_TEXTURE_IMAGE_UNITS];
 
    /** Buffer for saving the sampled texture colors.
     * Needed for GL_ARB_texture_env_crossbar implementation.