Major rework of tnl module

New array_cache module
Support 8 texture units in core mesa (now support 8 everywhere)
Rework core mesa statechange operations to avoid flushing on many
noop statechanges.

1 files changed, 355 insertions, 0 deletions

diff --git a/src/mesa/tnl/t_array_api.c b/src/mesa/tnl/t_array_api.c
new file mode 100644
index 00000000000..83d10337be3
--- /dev/null
+++ b/src/mesa/tnl/t_array_api.c
@@ -0,0 +1,355 @@
+/* $Id: t_array_api.c,v 1.1 2000/12/26 05:09:32 keithw Exp $ */
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.5
+ *
+ * Copyright (C) 1999-2000  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"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * 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]>
+ */
+
+#include "glheader.h"
+#include "api_validate.h"
+#include "context.h"
+#include "macros.h"
+#include "mmath.h"
+#include "mem.h"
+#include "state.h"
+#include "mtypes.h"
+
+#include "array_cache/acache.h"
+
+#include "t_array_api.h"
+#include "t_array_import.h"
+#include "t_imm_api.h"
+#include "t_imm_exec.h"
+#include "t_context.h"
+#include "t_pipeline.h"
+
+
+
+
+
+void
+_tnl_DrawArrays(GLenum mode, GLint start, GLsizei count)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   struct vertex_buffer *VB = &tnl->vb;
+
+   /* Check arguments, etc.
+    */
+   if (!_mesa_validate_DrawArrays( ctx, mode, start, count ))
+      return;
+
+   if (tnl->pipeline.build_state_changes)
+      _tnl_validate_pipeline( ctx );
+
+   if (!ctx->CompileFlag && count - start < ctx->Const.MaxArrayLockSize) {
+      FLUSH_CURRENT( ctx, 0 );
+
+      if (ctx->Array.LockCount)
+      {
+	 if (start < ctx->Array.LockFirst) start = ctx->Array.LockFirst;
+	 if (count > ctx->Array.LockCount) count = ctx->Array.LockCount;
+	 if (start >= count) return;
+
+	 /* Locked drawarrays.  Reuse any previously transformed data.
+	  */
+	 _tnl_vb_bind_arrays( ctx, ctx->Array.LockFirst, ctx->Array.LockCount );
+	 VB->FirstPrimitive = start;
+	 VB->Primitive[start] = mode | PRIM_BEGIN | PRIM_END | PRIM_LAST;
+	 VB->PrimitiveLength[start] = count - start;
+	 _tnl_run_pipeline( ctx );
+      } else {
+	 /* The arrays are small enough to fit in a single VB; just bind
+	  * them and go.  Any untransformed data will be copied on
+	  * clipping.
+	  * 
+	  * Invalidate any locked data dependent on these arrays.
+	  */
+	 _tnl_vb_bind_arrays( ctx, start, count );
+	 VB->FirstPrimitive = 0;
+	 VB->Primitive[0] = mode | PRIM_BEGIN | PRIM_END | PRIM_LAST;
+	 VB->PrimitiveLength[0] = count - start;	 
+	 tnl->pipeline.run_input_changes |= ctx->Array._Enabled;
+	 _tnl_run_pipeline( ctx );
+	 tnl->pipeline.run_input_changes |= ctx->Array._Enabled;
+      }
+   } 
+   else {
+      /* Need to produce immediate structs, either for compiling or
+       * because the array range is too large to process in a single
+       * VB.  In GL_EXECUTE mode, this introduces two redundant
+       * operations: producing the flag array and computing the orflag
+       * of the flag array. 
+       */
+#if 0
+      if (_tnl_hard_begin( ctx, mode )) {
+	 GLuint j;
+	 for (j = 0 ; j < count ; ) {
+	    struct immediate *IM = TNL_CURRENT_IM(ctx);
+	    GLuint nr = MIN2( IMM_MAXDATA - IM->Start, count - j );
+	    GLuint sf = IM->Flag[IM->Start];
+
+	    _tnl_fill_immediate_drawarrays( ctx, IM, j, j+nr );
+
+	    if (j == 0) IM->Flag[IM->Start] |= sf;
+
+	    IM->Count = IM->Start + nr;
+	    j += nr;
+
+	    if (j == count)
+	       _tnl_end( ctx );
+
+	    _tnl_flush_immediate( IM );
+	 }
+      }
+#else
+      /* Simple alternative to above code.
+       */
+/*        if (_tnl_hard_begin( ctx, mode ))  */
+      _tnl_begin(ctx,mode);
+      {
+	 GLuint i;
+	 for (i=start;i<count;i++) {
+	    _tnl_array_element( ctx, i );
+	 }
+	 _tnl_end( ctx );
+      }
+#endif
+   }
+}
+
+
+
+static void _tnl_draw_range_elements( GLcontext *ctx, GLenum mode, 
+				      GLuint start, GLuint end, 
+				      GLsizei count, const GLuint *indices )
+   
+{
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   FLUSH_CURRENT( ctx, 0 );
+   
+   _tnl_vb_bind_arrays( ctx, start, end );
+
+   tnl->vb.FirstPrimitive = 0;
+   tnl->vb.Primitive[0] = mode | PRIM_BEGIN | PRIM_END | PRIM_LAST;
+   tnl->vb.PrimitiveLength[0] = count;
+   tnl->vb.Elts = (GLuint *)indices;
+
+   if (ctx->Array.LockCount) 
+      _tnl_run_pipeline( ctx );
+   else {
+      /* Note that arrays may have changed before/after execution.
+       */
+      tnl->pipeline.run_input_changes |= ctx->Array._Enabled;
+      _tnl_run_pipeline( ctx );
+      tnl->pipeline.run_input_changes |= ctx->Array._Enabled;
+   }
+}
+
+
+
+
+static void _tnl_draw_elements( GLcontext *ctx, GLenum mode, GLsizei count, 
+				const GLuint *indices)
+{
+#if 1
+   /* Optimized code that fakes the effect of calling
+    * _tnl_array_element for each index in the list.
+    */
+   if (_tnl_hard_begin( ctx, mode )) {
+      GLuint i,j;
+      for (j = 0 ; j < count ; ) {
+	 struct immediate *IM = TNL_CURRENT_IM(ctx);
+	 GLuint start = IM->Start;
+	 GLuint nr = MIN2( IMM_MAXDATA - start, count - j ) + start;
+	 GLuint sf = IM->Flag[start];
+	 IM->FlushElt |= 1;
+
+	 for (i = start ; i < nr ; i++) {
+	    IM->Elt[i] = (GLuint) *indices++;
+	    IM->Flag[i] = VERT_ELT;
+	 }
+
+	 if (j == 0) IM->Flag[start] |= sf;
+
+	 IM->Count = nr;
+	 j += nr - start;
+
+	 if (j == count)
+	    _tnl_end( ctx );
+
+	 _tnl_flush_immediate( IM );
+      }
+   }
+#else
+   /* Simple version of the above code.
+    */
+   if (_tnl_hard_begin(ctx, mode)) {
+      GLuint i;
+      for (i = 0 ; i < count ; i++)
+	 _tnl_array_element( ctx, indices[i] );
+      _tnl_end( ctx );
+   }
+#endif
+}
+
+
+void
+_tnl_DrawRangeElements(GLenum mode,
+		       GLuint start, GLuint end,
+		       GLsizei count, GLenum type, const GLvoid *indices)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   GLuint *ui_indices;
+
+   /* Check arguments, etc.
+    */
+   if (!_mesa_validate_DrawRangeElements( ctx, mode, start, end, count, 
+                                          type, indices ))
+      return;
+
+   if (tnl->pipeline.build_state_changes)
+      _tnl_validate_pipeline( ctx );
+
+   ui_indices = (GLuint *)_ac_import_elements( ctx, GL_UNSIGNED_INT, 
+					       count, type, indices );
+            
+
+   if (ctx->Array.LockCount) {
+      /* Are the arrays already locked?  If so we currently have to look
+       * at the whole locked range.
+       */
+      if (start >= ctx->Array.LockFirst && end <= ctx->Array.LockCount)
+	 _tnl_draw_range_elements( ctx, mode, 
+				   ctx->Array.LockFirst, 
+				   ctx->Array.LockCount,
+				   count, ui_indices );
+      else {
+	 /* The spec says referencing elements outside the locked
+	  * range is undefined.  I'm going to make it a noop this time
+	  * round, maybe come up with something beter before 3.6.  
+	  *
+	  * May be able to get away with just setting LockCount==0,
+	  * though this raises the problems of dependent state.  May
+	  * have to call glUnlockArrays() directly?
+	  */
+	 gl_problem( ctx, 
+		     "DrawRangeElements references "
+		     "elements outside locked range.");
+      }
+   }
+   else if (end - start < ctx->Const.MaxArrayLockSize) {
+      /* The arrays aren't locked but we can still fit them inside a single
+       * vertexbuffer.
+       */
+      _tnl_draw_range_elements( ctx, mode, start, end, count, ui_indices );
+   } else {
+      /* Range is too big to optimize:
+       */
+      _tnl_draw_elements( ctx, mode, count, ui_indices );
+   }
+}
+
+
+
+void
+_tnl_DrawElements(GLenum mode, GLsizei count, GLenum type, 
+		  const GLvoid *indices)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   GLuint *ui_indices;
+
+   /* Check arguments, etc.
+    */
+   if (!_mesa_validate_DrawElements( ctx, mode, count, type, indices ))
+      return;
+
+   if (tnl->pipeline.build_state_changes)
+      _tnl_validate_pipeline( ctx );
+
+   ui_indices = (GLuint *)_ac_import_elements( ctx, GL_UNSIGNED_INT, 
+					       count, type, indices );
+            
+   if (ctx->Array.LockCount) {
+      _tnl_draw_range_elements( ctx, mode, 
+				ctx->Array.LockFirst,
+				ctx->Array.LockCount,
+				count, ui_indices );
+   } 
+   else {
+      /* Scan the index list and see if we can use the locked path anyway.
+       */
+      GLuint max_elt = 0;
+      GLuint i;
+
+      for (i = 0 ; i < count ; i++) 
+	 if (ui_indices[i] > max_elt) max_elt = ui_indices[i];
+
+      if (max_elt < ctx->Const.MaxArrayLockSize && /* can we use it? */
+	  max_elt < count)	                   /* do we want to use it? */
+	 _tnl_draw_range_elements( ctx, mode, 0, max_elt, count, ui_indices );
+      else
+	 _tnl_draw_elements( ctx, mode, count, ui_indices );
+   }
+}
+
+
+void _tnl_array_init( GLcontext *ctx )
+{
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   struct vertex_arrays *tmp = &tnl->array_inputs;
+   GLvertexformat *vfmt = &(TNL_CONTEXT(ctx)->vtxfmt);
+   GLuint i;
+
+   vfmt->DrawArrays = _tnl_DrawArrays;
+   vfmt->DrawElements = _tnl_DrawElements;
+   vfmt->DrawRangeElements = _tnl_DrawRangeElements;
+
+   /* Setup vector pointers that will be used to bind arrays to VB's.
+    */
+   gl_vector4f_init( &tmp->Obj, 0, 0 );
+   gl_vector3f_init( &tmp->Normal, 0, 0 );
+   gl_vector4ub_init( &tmp->Color, 0, 0 );
+   gl_vector4ub_init( &tmp->SecondaryColor, 0, 0 );
+   gl_vector1f_init( &tmp->FogCoord, 0, 0 );
+   gl_vector1ui_init( &tmp->Index, 0, 0 );
+   gl_vector1ub_init( &tmp->EdgeFlag, 0, 0 );
+
+   for (i = 0; i < ctx->Const.MaxTextureUnits; i++) 
+      gl_vector4f_init( &tmp->TexCoord[i], 0, 0);
+
+   tnl->tmp_primitive = (GLuint *)MALLOC(sizeof(GLuint)*tnl->vb.Size);
+   tnl->tmp_primitive_length = (GLuint *)MALLOC(sizeof(GLuint)*tnl->vb.Size);
+}
+
+
+void _tnl_array_destroy( GLcontext *ctx )
+{
+   TNLcontext *tnl = TNL_CONTEXT(ctx);
+   if (tnl->tmp_primitive_length) FREE(tnl->tmp_primitive_length);
+   if (tnl->tmp_primitive) FREE(tnl->tmp_primitive);
+}