r300_swtcl.c

Mesa is an open-source implementation of the OpenGL specification - a system for rendering interacti

/**************************************************************************

Copyright (C) 2007 Dave Airlie

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
on the rights to use, copy, modify, merge, publish, distribute, sub
license, 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 (including the next
paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL
THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS 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:
 *   Dave Airlie <[e-mail]>
 */

/* derived from r200 swtcl path */



#include "glheader.h"
#include "mtypes.h"
#include "colormac.h"
#include "enums.h"
#include "image.h"
#include "imports.h"
#include "light.h"
#include "macros.h"

#include "swrast/s_context.h"
#include "swrast/s_fog.h"
#include "swrast_setup/swrast_setup.h"
#include "math/m_translate.h"
#include "tnl/tnl.h"
#include "tnl/t_context.h"
#include "tnl/t_pipeline.h"

#include "r300_context.h"
#include "r300_swtcl.h"
#include "r300_state.h"
#include "r300_ioctl.h"
#include "r300_emit.h"
#include "r300_mem.h"

static void flush_last_swtcl_prim( r300ContextPtr rmesa  );


void r300EmitVertexAOS(r300ContextPtr rmesa, GLuint vertex_size, GLuint offset);
void r300EmitVbufPrim(r300ContextPtr rmesa, GLuint primitive, GLuint vertex_nr);
#define EMIT_ATTR( ATTR, STYLE )					\
do {									\
   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].attrib = (ATTR);	\
   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].format = (STYLE);	\
   rmesa->swtcl.vertex_attr_count++;					\
} while (0)

#define EMIT_PAD( N )							\
do {									\
   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].attrib = 0;		\
   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].format = EMIT_PAD;	\
   rmesa->swtcl.vertex_attrs[rmesa->swtcl.vertex_attr_count].offset = (N);		\
   rmesa->swtcl.vertex_attr_count++;					\
} while (0)

static void r300SetVertexFormat( GLcontext *ctx )
{
	r300ContextPtr rmesa = R300_CONTEXT( ctx );
	TNLcontext *tnl = TNL_CONTEXT(ctx);
	struct vertex_buffer *VB = &tnl->vb;
	DECLARE_RENDERINPUTS(index_bitset);
	GLuint InputsRead = 0, OutputsWritten = 0;
	int vap_fmt_0 = 0;
	int vap_vte_cntl = 0;
	int offset = 0;
	int vte = 0;
	GLint inputs[VERT_ATTRIB_MAX];
	GLint tab[VERT_ATTRIB_MAX];
	int swizzle[VERT_ATTRIB_MAX][4];
	GLuint i, nr;
	GLuint sz, vap_fmt_1 = 0;

	DECLARE_RENDERINPUTS(render_inputs_bitset);
	RENDERINPUTS_COPY(render_inputs_bitset, tnl->render_inputs_bitset);
	RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
	RENDERINPUTS_COPY(rmesa->state.render_inputs_bitset, render_inputs_bitset);

	vte = rmesa->hw.vte.cmd[1];
	vte &= ~(R300_VTX_XY_FMT | R300_VTX_Z_FMT | R300_VTX_W0_FMT);
	/* Important:
	 */
	if ( VB->NdcPtr != NULL ) {
		VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
		vte |= R300_VTX_XY_FMT | R300_VTX_Z_FMT;
	}
	else {
		VB->AttribPtr[VERT_ATTRIB_POS] = VB->ClipPtr;
		vte |= R300_VTX_W0_FMT;
	}

	assert( VB->AttribPtr[VERT_ATTRIB_POS] != NULL );
	rmesa->swtcl.vertex_attr_count = 0;

	/* EMIT_ATTR's must be in order as they tell t_vertex.c how to
	 * build up a hardware vertex.
	 */
	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POS)) {
		sz = VB->AttribPtr[VERT_ATTRIB_POS]->size;
		InputsRead |= 1 << VERT_ATTRIB_POS;
		OutputsWritten |= 1 << VERT_RESULT_HPOS;
		EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_1F + sz - 1 );
		offset = sz;
	} else {
		offset = 4;
		EMIT_PAD(4 * sizeof(float));
	}

	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POINTSIZE )) {
		EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F );
		vap_fmt_0 |=  R300_VAP_OUTPUT_VTX_FMT_0__PT_SIZE_PRESENT;
		offset += 1;
	}

	if (RENDERINPUTS_TEST(index_bitset, _TNL_ATTRIB_COLOR0)) {
		sz = VB->AttribPtr[VERT_ATTRIB_COLOR0]->size;
	        rmesa->swtcl.coloroffset = offset;
		InputsRead |= 1 << VERT_ATTRIB_COLOR0;
		OutputsWritten |= 1 << VERT_RESULT_COL0;
		EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_1F + sz - 1 );
		offset += sz;
	}

	rmesa->swtcl.specoffset = 0;
	if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
		sz = VB->AttribPtr[VERT_ATTRIB_COLOR1]->size;
		rmesa->swtcl.specoffset = offset;
		EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_1F + sz - 1 );
		InputsRead |= 1 << VERT_ATTRIB_COLOR1;
		OutputsWritten |= 1 << VERT_RESULT_COL1;
	}

	if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
		int i;

		for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
			if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
				sz = VB->TexCoordPtr[i]->size;
				InputsRead |= 1 << (VERT_ATTRIB_TEX0 + i);
				OutputsWritten |= 1 << (VERT_RESULT_TEX0 + i);
				EMIT_ATTR( _TNL_ATTRIB_TEX0+i, EMIT_1F + sz - 1 );
				vap_fmt_1 |= sz << (3 * i);
			}
		}
	}

	for (i = 0, nr = 0; i < VERT_ATTRIB_MAX; i++) {
		if (InputsRead & (1 << i)) {
			inputs[i] = nr++;
		} else {
			inputs[i] = -1;
		}
	}
	
	/* Fixed, apply to vir0 only */
	if (InputsRead & (1 << VERT_ATTRIB_POS))
		inputs[VERT_ATTRIB_POS] = 0;
	if (InputsRead & (1 << VERT_ATTRIB_COLOR0))
		inputs[VERT_ATTRIB_COLOR0] = 2;
	if (InputsRead & (1 << VERT_ATTRIB_COLOR1))
		inputs[VERT_ATTRIB_COLOR1] = 3;
	for (i = VERT_ATTRIB_TEX0; i <= VERT_ATTRIB_TEX7; i++)
		if (InputsRead & (1 << i))
			inputs[i] = 6 + (i - VERT_ATTRIB_TEX0);
	
	for (i = 0, nr = 0; i < VERT_ATTRIB_MAX; i++) {
		if (InputsRead & (1 << i)) {
			tab[nr++] = i;
		}
	}
	
	for (i = 0; i < nr; i++) {
		int ci;
		
		swizzle[i][0] = SWIZZLE_ZERO;
		swizzle[i][1] = SWIZZLE_ZERO;
		swizzle[i][2] = SWIZZLE_ZERO;
		swizzle[i][3] = SWIZZLE_ONE;

		for (ci = 0; ci < VB->AttribPtr[tab[i]]->size; ci++) {
			swizzle[i][ci] = ci;
		}
	}

	R300_NEWPRIM(rmesa);
	R300_STATECHANGE(rmesa, vir[0]);
	((drm_r300_cmd_header_t *) rmesa->hw.vir[0].cmd)->packet0.count =
		r300VAPInputRoute0(&rmesa->hw.vir[0].cmd[R300_VIR_CNTL_0],
				   VB->AttribPtr, inputs, tab, nr);
	R300_STATECHANGE(rmesa, vir[1]);
	((drm_r300_cmd_header_t *) rmesa->hw.vir[1].cmd)->packet0.count =
		r300VAPInputRoute1(&rmesa->hw.vir[1].cmd[R300_VIR_CNTL_0], swizzle,
				   nr);
   
	R300_STATECHANGE(rmesa, vic);
	rmesa->hw.vic.cmd[R300_VIC_CNTL_0] = r300VAPInputCntl0(ctx, InputsRead);
	rmesa->hw.vic.cmd[R300_VIC_CNTL_1] = r300VAPInputCntl1(ctx, InputsRead);
   
	R300_STATECHANGE(rmesa, vof);
	rmesa->hw.vof.cmd[R300_VOF_CNTL_0] = r300VAPOutputCntl0(ctx, OutputsWritten);
	rmesa->hw.vof.cmd[R300_VOF_CNTL_1] = vap_fmt_1;
   
	rmesa->swtcl.vertex_size =
		_tnl_install_attrs( ctx,
				    rmesa->swtcl.vertex_attrs, 
				    rmesa->swtcl.vertex_attr_count,
				    NULL, 0 );
	
	rmesa->swtcl.vertex_size /= 4;

	RENDERINPUTS_COPY( rmesa->tnl_index_bitset, index_bitset );


	R300_STATECHANGE(rmesa, vte);
	rmesa->hw.vte.cmd[1] = vte;
	rmesa->hw.vte.cmd[2] = rmesa->swtcl.vertex_size;
}


/* Flush vertices in the current dma region.
 */
static void flush_last_swtcl_prim( r300ContextPtr rmesa  )
{
	if (RADEON_DEBUG & DEBUG_IOCTL)
		fprintf(stderr, "%s\n", __FUNCTION__);
	
	rmesa->dma.flush = NULL;

	if (rmesa->dma.current.buf) {
		struct r300_dma_region *current = &rmesa->dma.current;
		GLuint current_offset = GET_START(current);

		assert (current->start + 
			rmesa->swtcl.numverts * rmesa->swtcl.vertex_size * 4 ==
			current->ptr);

		if (rmesa->dma.current.start != rmesa->dma.current.ptr) {

			r300EnsureCmdBufSpace( rmesa, rmesa->hw.max_state_size + (12*sizeof(int)), __FUNCTION__);
			
			r300EmitState(rmesa);
			
			r300EmitVertexAOS( rmesa,
					   rmesa->swtcl.vertex_size,
					   current_offset);
			
			r300EmitVbufPrim( rmesa,
					  rmesa->swtcl.hw_primitive,
					  rmesa->swtcl.numverts);
			
			r300EmitCacheFlush(rmesa);
		}
		
		rmesa->swtcl.numverts = 0;
		current->start = current->ptr;
	}
}

/* Alloc space in the current dma region.
 */
static void *
r300AllocDmaLowVerts( r300ContextPtr rmesa, int nverts, int vsize )
{
	GLuint bytes = vsize * nverts;

	if ( rmesa->dma.current.ptr + bytes > rmesa->dma.current.end ) 
		r300RefillCurrentDmaRegion( rmesa, bytes);

	if (!rmesa->dma.flush) {
		rmesa->radeon.glCtx->Driver.NeedFlush |= FLUSH_STORED_VERTICES;
		rmesa->dma.flush = flush_last_swtcl_prim;
	}

	ASSERT( vsize == rmesa->swtcl.vertex_size * 4 );
	ASSERT( rmesa->dma.flush == flush_last_swtcl_prim );
	ASSERT( rmesa->dma.current.start + 
		rmesa->swtcl.numverts * rmesa->swtcl.vertex_size * 4 ==
		rmesa->dma.current.ptr );

	{
		GLubyte *head = (GLubyte *) (rmesa->dma.current.address + rmesa->dma.current.ptr);
		rmesa->dma.current.ptr += bytes;
		rmesa->swtcl.numverts += nverts;
		return head;
	}
}

static GLuint reduced_prim[] = {
  GL_POINTS,
  GL_LINES,
  GL_LINES,
  GL_LINES,
  GL_TRIANGLES,
  GL_TRIANGLES,
  GL_TRIANGLES,
  GL_TRIANGLES,
  GL_TRIANGLES,
  GL_TRIANGLES,
};

static void r300RasterPrimitive( GLcontext *ctx, GLuint prim );
static void r300RenderPrimitive( GLcontext *ctx, GLenum prim );
//static void r300ResetLineStipple( GLcontext *ctx );

/***********************************************************************
 *                    Emit primitives as inline vertices               *
 ***********************************************************************/


#define HAVE_POINTS      1
#define HAVE_LINES       1
#define HAVE_LINE_STRIPS 1
#define HAVE_TRIANGLES   1
#define HAVE_TRI_STRIPS  1
#define HAVE_TRI_STRIP_1 0
#define HAVE_TRI_FANS    1
#define HAVE_QUADS       0
#define HAVE_QUAD_STRIPS 0
#define HAVE_POLYGONS    1
#define HAVE_ELTS        1

#undef LOCAL_VARS
#undef ALLOC_VERTS
#define CTX_ARG r300ContextPtr rmesa
#define GET_VERTEX_DWORDS() rmesa->swtcl.vertex_size