/************************************************************************** * * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. * 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, 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 TUNGSTEN GRAPHICS 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: * Zack Rusin * Keith Whitwell * Brian Paul */ #include "pipe/p_util.h" #include "pipe/p_shader_tokens.h" #include "draw_private.h" #include "draw_context.h" #include "draw_vs.h" #include "tgsi/util/tgsi_parse.h" #ifdef MESA_LLVM #include "gallivm/gallivm.h" struct draw_llvm_vertex_shader { struct draw_vertex_shader base; struct gallivm_prog *llvm_prog; }; static void vs_llvm_prepare( struct draw_vertex_shader *base, struct draw_context *draw ) { draw_update_vertex_fetch( draw ); } /** * Transform vertices with the current vertex program/shader * Up to four vertices can be shaded at a time. * \param vbuffer the input vertex data * \param elts indexes of four input vertices * \param count number of vertices to shade [1..4] * \param vOut array of pointers to four output vertices */ static boolean vs_llvm_run( struct draw_vertex_shader *base, struct draw_context *draw, const unsigned *elts, unsigned count, void *vOut ) { struct draw_llvm_vertex_shader *shader = (struct draw_llvm_vertex_shader *)base; struct tgsi_exec_machine *machine = &draw->machine; unsigned int j; unsigned int clipped = 0; ALIGN16_DECL(struct tgsi_exec_vector, inputs, PIPE_MAX_ATTRIBS); ALIGN16_DECL(struct tgsi_exec_vector, outputs, PIPE_MAX_ATTRIBS); const float *scale = draw->viewport.scale; const float *trans = draw->viewport.translate; assert(count <= 4); assert(draw->vertex_shader->state->output_semantic_name[0] == TGSI_SEMANTIC_POSITION); /* Consts does not require 16 byte alignment. */ machine->Consts = (float (*)[4]) draw->user.constants; machine->Inputs = ALIGN16_ASSIGN(inputs); if (draw->rasterizer->bypass_vs) { /* outputs are just the inputs */ machine->Outputs = machine->Inputs; } else { machine->Outputs = ALIGN16_ASSIGN(outputs); } draw->vertex_fetch.fetch_func( draw, machine, elts, count ); if (!draw->rasterizer->bypass_vs) { /* run shader */ gallivm_cpu_vs_exec(shader->llvm_prog, machine->Inputs, machine->Outputs, machine->Consts, machine->Temps); } /* store machine results */ for (j = 0; j < count; j++) { unsigned slot; float x, y, z, w; x = vOut[j]->clip[0] = machine->Outputs[0].xyzw[0].f[j]; y = vOut[j]->clip[1] = machine->Outputs[0].xyzw[1].f[j]; z = vOut[j]->clip[2] = machine->Outputs[0].xyzw[2].f[j]; w = vOut[j]->clip[3] = machine->Outputs[0].xyzw[3].f[j]; if (!draw->rasterizer->bypass_clipping) { vOut[j]->clipmask = compute_clipmask(vOut[j]->clip, draw->plane, draw->nr_planes); clipped += vOut[j]->clipmask; /* divide by w */ w = 1.0f / w; x *= w; y *= w; z *= w; } else { vOut[j]->clipmask = 0; } vOut[j]->edgeflag = 1; vOut[j]->vertex_id = UNDEFINED_VERTEX_ID; if (!draw->identity_viewport) { /* Viewport mapping */ vOut[j]->data[0][0] = x * scale[0] + trans[0]; vOut[j]->data[0][1] = y * scale[1] + trans[1]; vOut[j]->data[0][2] = z * scale[2] + trans[2]; vOut[j]->data[0][3] = w; } else { vOut[j]->data[0][0] = x; vOut[j]->data[0][1] = y; vOut[j]->data[0][2] = z; vOut[j]->data[0][3] = w; } /* Remaining attributes are packed into sequential post-transform * vertex attrib slots. */ for (slot = 1; slot < draw->num_vs_outputs; slot++) { vOut[j]->data[slot][0] = machine->Outputs[slot].xyzw[0].f[j]; vOut[j]->data[slot][1] = machine->Outputs[slot].xyzw[1].f[j]; vOut[j]->data[slot][2] = machine->Outputs[slot].xyzw[2].f[j]; vOut[j]->data[slot][3] = machine->Outputs[slot].xyzw[3].f[j]; } } /* loop over vertices */ return clipped != 0; } static void vs_llvm_delete( struct draw_vertex_shader *base ) { struct draw_llvm_vertex_shader *shader = (struct draw_llvm_vertex_shader *)base; /* Do something to free compiled shader: */ FREE( (void*) shader->base.state.tokens ); FREE( shader ); } struct draw_vertex_shader * draw_create_vs_llvm(struct draw_context *draw, const struct pipe_shader_state *templ) { struct draw_llvm_vertex_shader *vs; uint nt = tgsi_num_tokens(templ->tokens); vs = CALLOC_STRUCT( draw_llvm_vertex_shader ); if (vs == NULL) return NULL; /* we make a private copy of the tokens */ vs->base.state.tokens = mem_dup(templ->tokens, nt * sizeof(templ->tokens[0])); vs->base.prepare = vs_llvm_prepare; vs->base.run = vs_llvm_run; vs->base.delete = vs_llvm_delete; { struct gallivm_ir *ir = gallivm_ir_new(GALLIVM_VS); gallivm_ir_set_layout(ir, GALLIVM_SOA); gallivm_ir_set_components(ir, 4); gallivm_ir_fill_from_tgsi(ir, vs->base.state->tokens); vs->llvm_prog = gallivm_ir_compile(ir); gallivm_ir_delete(ir); } draw->engine = gallivm_global_cpu_engine(); /* XXX: Why are there two versions of this? Shouldn't creating the * engine be a separate operation to compiling a shader? */ if (!draw->engine) { draw->engine = gallivm_cpu_engine_create(vs->llvm_prog); } else { gallivm_cpu_jit_compile(draw->engine, vs->llvm_prog); } return &vs->base; } #else struct draw_vertex_shader * draw_create_vs_llvm(struct draw_context *draw, const struct pipe_shader_state *shader) { return NULL; } #endif