/************************************************************************** * * 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: * Keith Whitwell */ #include "pipe/p_context.h" #include "util/u_memory.h" #include "util/u_math.h" #include "util/u_cpu_detect.h" #include "util/u_inlines.h" #include "util/u_helpers.h" #include "draw_context.h" #include "draw_vs.h" #include "draw_gs.h" #if HAVE_LLVM #include "gallivm/lp_bld_init.h" #include "gallivm/lp_bld_limits.h" #include "draw_llvm.h" boolean draw_get_option_use_llvm(void) { static boolean first = TRUE; static boolean value; if (first) { first = FALSE; value = debug_get_bool_option("DRAW_USE_LLVM", TRUE); #ifdef PIPE_ARCH_X86 util_cpu_detect(); /* require SSE2 due to LLVM PR6960. */ if (!util_cpu_caps.has_sse2) value = FALSE; #endif } return value; } #endif /** * Create new draw module context with gallivm state for LLVM JIT. */ static struct draw_context * draw_create_context(struct pipe_context *pipe, boolean try_llvm) { struct draw_context *draw = CALLOC_STRUCT( draw_context ); if (draw == NULL) goto err_out; #if HAVE_LLVM if (try_llvm && draw_get_option_use_llvm()) { draw->llvm = draw_llvm_create(draw); if (!draw->llvm) goto err_destroy; } #endif draw->pipe = pipe; if (!draw_init(draw)) goto err_destroy; return draw; err_destroy: draw_destroy( draw ); err_out: return NULL; } /** * Create new draw module context, with LLVM JIT. */ struct draw_context * draw_create(struct pipe_context *pipe) { return draw_create_context(pipe, TRUE); } /** * Create a new draw context, without LLVM JIT. */ struct draw_context * draw_create_no_llvm(struct pipe_context *pipe) { return draw_create_context(pipe, FALSE); } boolean draw_init(struct draw_context *draw) { /* * Note that several functions compute the clipmask of the predefined * formats with hardcoded formulas instead of using these. So modifications * here must be reflected there too. */ ASSIGN_4V( draw->plane[0], -1, 0, 0, 1 ); ASSIGN_4V( draw->plane[1], 1, 0, 0, 1 ); ASSIGN_4V( draw->plane[2], 0, -1, 0, 1 ); ASSIGN_4V( draw->plane[3], 0, 1, 0, 1 ); ASSIGN_4V( draw->plane[4], 0, 0, 1, 1 ); /* yes these are correct */ ASSIGN_4V( draw->plane[5], 0, 0, -1, 1 ); /* mesa's a bit wonky */ draw->clip_xy = TRUE; draw->clip_z = TRUE; draw->pt.user.planes = (float (*) [DRAW_TOTAL_CLIP_PLANES][4]) &(draw->plane[0]); if (!draw_pipeline_init( draw )) return FALSE; if (!draw_pt_init( draw )) return FALSE; if (!draw_vs_init( draw )) return FALSE; if (!draw_gs_init( draw )) return FALSE; draw->quads_always_flatshade_last = !draw->pipe->screen->get_param( draw->pipe->screen, PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION); return TRUE; } /* * Called whenever we're starting to draw a new instance. * Some internal structures don't want to have to reset internal * members on each invocation (because their state might have to persist * between multiple primitive restart rendering call) but might have to * for each new instance. * This is particularly the case for primitive id's in geometry shader. */ void draw_new_instance(struct draw_context *draw) { draw_geometry_shader_new_instance(draw->gs.geometry_shader); } void draw_destroy( struct draw_context *draw ) { struct pipe_context *pipe; unsigned i, j; if (!draw) return; pipe = draw->pipe; /* free any rasterizer CSOs that we may have created. */ for (i = 0; i < 2; i++) { for (j = 0; j < 2; j++) { if (draw->rasterizer_no_cull[i][j]) { pipe->delete_rasterizer_state(pipe, draw->rasterizer_no_cull[i][j]); } } } for (i = 0; i < draw->pt.nr_vertex_buffers; i++) { pipe_resource_reference(&draw->pt.vertex_buffer[i].buffer, NULL); } /* Not so fast -- we're just borrowing this at the moment. * if (draw->render) draw->render->destroy( draw->render ); */ draw_pipeline_destroy( draw ); draw_pt_destroy( draw ); draw_vs_destroy( draw ); draw_gs_destroy( draw ); #ifdef HAVE_LLVM if (draw->llvm) draw_llvm_destroy( draw->llvm ); #endif FREE( draw ); } void draw_flush( struct draw_context *draw ) { draw_do_flush( draw, DRAW_FLUSH_BACKEND ); } /** * Specify the Minimum Resolvable Depth factor for polygon offset. * This factor potentially depends on the number of Z buffer bits, * the rasterization algorithm and the arithmetic performed on Z * values between vertex shading and rasterization. It will vary * from one driver to another. */ void draw_set_mrd(struct draw_context *draw, double mrd) { draw->mrd = mrd; } static void update_clip_flags( struct draw_context *draw ) { draw->clip_xy = !draw->driver.bypass_clip_xy; draw->guard_band_xy = (!draw->driver.bypass_clip_xy && draw->driver.guard_band_xy); draw->clip_z = (!draw->driver.bypass_clip_z && draw->rasterizer && draw->rasterizer->depth_clip); draw->clip_user = draw->rasterizer && draw->rasterizer->clip_plane_enable != 0; } /** * Register new primitive rasterization/rendering state. * This causes the drawing pipeline to be rebuilt. */ void draw_set_rasterizer_state( struct draw_context *draw, const struct pipe_rasterizer_state *raster, void *rast_handle ) { if (!draw->suspend_flushing) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->rasterizer = raster; draw->rast_handle = rast_handle; update_clip_flags(draw); } } /* With a little more work, llvmpipe will be able to turn this off and * do its own x/y clipping. * * Some hardware can turn off clipping altogether - in particular any * hardware with a TNL unit can do its own clipping, even if it is * relying on the draw module for some other reason. */ void draw_set_driver_clipping( struct draw_context *draw, boolean bypass_clip_xy, boolean bypass_clip_z, boolean guard_band_xy) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->driver.bypass_clip_xy = bypass_clip_xy; draw->driver.bypass_clip_z = bypass_clip_z; draw->driver.guard_band_xy = guard_band_xy; update_clip_flags(draw); } /** * Plug in the primitive rendering/rasterization stage (which is the last * stage in the drawing pipeline). * This is provided by the device driver. */ void draw_set_rasterize_stage( struct draw_context *draw, struct draw_stage *stage ) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.rasterize = stage; } /** * Set the draw module's clipping state. */ void draw_set_clip_state( struct draw_context *draw, const struct pipe_clip_state *clip ) { draw_do_flush(draw, DRAW_FLUSH_PARAMETER_CHANGE); memcpy(&draw->plane[6], clip->ucp, sizeof(clip->ucp)); } /** * Set the draw module's viewport state. */ void draw_set_viewport_state( struct draw_context *draw, const struct pipe_viewport_state *viewport ) { draw_do_flush(draw, DRAW_FLUSH_PARAMETER_CHANGE); draw->viewport = *viewport; /* struct copy */ draw->identity_viewport = (viewport->scale[0] == 1.0f && viewport->scale[1] == 1.0f && viewport->scale[2] == 1.0f && viewport->scale[3] == 1.0f && viewport->translate[0] == 0.0f && viewport->translate[1] == 0.0f && viewport->translate[2] == 0.0f && viewport->translate[3] == 0.0f); draw_vs_set_viewport( draw, viewport ); } void draw_set_vertex_buffers(struct draw_context *draw, unsigned start_slot, unsigned count, const struct pipe_vertex_buffer *buffers) { assert(start_slot + count <= PIPE_MAX_ATTRIBS); util_set_vertex_buffers_count(draw->pt.vertex_buffer, &draw->pt.nr_vertex_buffers, buffers, start_slot, count); } void draw_set_vertex_elements(struct draw_context *draw, unsigned count, const struct pipe_vertex_element *elements) { assert(count <= PIPE_MAX_ATTRIBS); /* We could improve this by only flushing the frontend and the fetch part * of the middle. This would avoid recalculating the emit keys.*/ draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); memcpy(draw->pt.vertex_element, elements, count * sizeof(elements[0])); draw->pt.nr_vertex_elements = count; } /** * Tell drawing context where to find mapped vertex buffers. */ void draw_set_mapped_vertex_buffer(struct draw_context *draw, unsigned attr, const void *buffer) { draw->pt.user.vbuffer[attr] = buffer; } void draw_set_mapped_constant_buffer(struct draw_context *draw, unsigned shader_type, unsigned slot, const void *buffer, unsigned size ) { debug_assert(shader_type == PIPE_SHADER_VERTEX || shader_type == PIPE_SHADER_GEOMETRY); debug_assert(slot < PIPE_MAX_CONSTANT_BUFFERS); draw_do_flush(draw, DRAW_FLUSH_PARAMETER_CHANGE); switch (shader_type) { case PIPE_SHADER_VERTEX: draw->pt.user.vs_constants[slot] = buffer; draw->pt.user.vs_constants_size[slot] = size; break; case PIPE_SHADER_GEOMETRY: draw->pt.user.gs_constants[slot] = buffer; draw->pt.user.gs_constants_size[slot] = size; break; default: assert(0 && "invalid shader type in draw_set_mapped_constant_buffer"); } } /** * Tells the draw module to draw points with triangles if their size * is greater than this threshold. */ void draw_wide_point_threshold(struct draw_context *draw, float threshold) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.wide_point_threshold = threshold; } /** * Should the draw module handle point->quad conversion for drawing sprites? */ void draw_wide_point_sprites(struct draw_context *draw, boolean draw_sprite) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.wide_point_sprites = draw_sprite; } /** * Tells the draw module to draw lines with triangles if their width * is greater than this threshold. */ void draw_wide_line_threshold(struct draw_context *draw, float threshold) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.wide_line_threshold = roundf(threshold); } /** * Tells the draw module whether or not to implement line stipple. */ void draw_enable_line_stipple(struct draw_context *draw, boolean enable) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.line_stipple = enable; } /** * Tells draw module whether to convert points to quads for sprite mode. */ void draw_enable_point_sprites(struct draw_context *draw, boolean enable) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->pipeline.point_sprite = enable; } void draw_set_force_passthrough( struct draw_context *draw, boolean enable ) { draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); draw->force_passthrough = enable; } /** * Allocate an extra vertex/geometry shader vertex attribute, if it doesn't * exist already. * * This is used by some of the optional draw module stages such * as wide_point which may need to allocate additional generic/texcoord * attributes. */ int draw_alloc_extra_vertex_attrib(struct draw_context *draw, uint semantic_name, uint semantic_index) { int slot; uint num_outputs; uint n; slot = draw_find_shader_output(draw, semantic_name, semantic_index); if (slot > 0) { return slot; } num_outputs = draw_current_shader_outputs(draw); n = draw->extra_shader_outputs.num; assert(n < Elements(draw->extra_shader_outputs.semantic_name)); draw->extra_shader_outputs.semantic_name[n] = semantic_name; draw->extra_shader_outputs.semantic_index[n] = semantic_index; draw->extra_shader_outputs.slot[n] = num_outputs + n; draw->extra_shader_outputs.num++; return draw->extra_shader_outputs.slot[n]; } /** * Remove all extra vertex attributes that were allocated with * draw_alloc_extra_vertex_attrib(). */ void draw_remove_extra_vertex_attribs(struct draw_context *draw) { draw->extra_shader_outputs.num = 0; } /** * If a geometry shader is present, return its info, else the vertex shader's * info. */ struct tgsi_shader_info * draw_get_shader_info(const struct draw_context *draw) { if (draw->gs.geometry_shader) { return &draw->gs.geometry_shader->info; } else { return &draw->vs.vertex_shader->info; } } /** * Ask the draw module for the location/slot of the given vertex attribute in * a post-transformed vertex. * * With this function, drivers that use the draw module should have no reason * to track the current vertex/geometry shader. * * Note that the draw module may sometimes generate vertices with extra * attributes (such as texcoords for AA lines). The driver can call this * function to find those attributes. * * Zero is returned if the attribute is not found since this is * a don't care / undefined situtation. Returning -1 would be a bit more * work for the drivers. */ int draw_find_shader_output(const struct draw_context *draw, uint semantic_name, uint semantic_index) { const struct tgsi_shader_info *info = draw_get_shader_info(draw); uint i; for (i = 0; i < info->num_outputs; i++) { if (info->output_semantic_name[i] == semantic_name && info->output_semantic_index[i] == semantic_index) return i; } /* Search the extra vertex attributes */ for (i = 0; i < draw->extra_shader_outputs.num; i++) { if (draw->extra_shader_outputs.semantic_name[i] == semantic_name && draw->extra_shader_outputs.semantic_index[i] == semantic_index) { return draw->extra_shader_outputs.slot[i]; } } return 0; } /** * Return total number of the shader outputs. This function is similar to * draw_current_shader_outputs() but this function also counts any extra * vertex/geometry output attributes that may be filled in by some draw * stages (such as AA point, AA line). * * If geometry shader is present, its output will be returned, * if not vertex shader is used. */ uint draw_num_shader_outputs(const struct draw_context *draw) { const struct tgsi_shader_info *info = draw_get_shader_info(draw); uint count; count = info->num_outputs; count += draw->extra_shader_outputs.num; return count; } /** * Provide TGSI sampler objects for vertex/geometry shaders that use * texture fetches. This state only needs to be set once per context. * This might only be used by software drivers for the time being. */ void draw_texture_sampler(struct draw_context *draw, uint shader, struct tgsi_sampler *sampler) { if (shader == PIPE_SHADER_VERTEX) { draw->vs.tgsi.sampler = sampler; } else { debug_assert(shader == PIPE_SHADER_GEOMETRY); draw->gs.tgsi.sampler = sampler; } } void draw_set_render( struct draw_context *draw, struct vbuf_render *render ) { draw->render = render; } /** * Tell the draw module where vertex indexes/elements are located, and * their size (in bytes). * * Note: the caller must apply the pipe_index_buffer::offset value to * the address. The draw module doesn't do that. */ void draw_set_indexes(struct draw_context *draw, const void *elements, unsigned elem_size) { assert(elem_size == 0 || elem_size == 1 || elem_size == 2 || elem_size == 4); draw->pt.user.elts = elements; draw->pt.user.eltSizeIB = elem_size; } /* Revamp me please: */ void draw_do_flush( struct draw_context *draw, unsigned flags ) { if (!draw->suspend_flushing) { assert(!draw->flushing); /* catch inadvertant recursion */ draw->flushing = TRUE; draw_pipeline_flush( draw, flags ); draw_pt_flush( draw, flags ); draw->flushing = FALSE; } } /** * Return the number of output attributes produced by the geometry * shader, if present. If no geometry shader, return the number of * outputs from the vertex shader. * \sa draw_num_shader_outputs */ uint draw_current_shader_outputs(const struct draw_context *draw) { if (draw->gs.geometry_shader) return draw->gs.num_gs_outputs; return draw->vs.num_vs_outputs; } /** * Return the index of the shader output which will contain the * vertex position. */ uint draw_current_shader_position_output(const struct draw_context *draw) { if (draw->gs.geometry_shader) return draw->gs.position_output; return draw->vs.position_output; } /** * Return the index of the shader output which will contain the * vertex position. */ uint draw_current_shader_clipvertex_output(const struct draw_context *draw) { return draw->vs.clipvertex_output; } uint draw_current_shader_clipdistance_output(const struct draw_context *draw, int index) { return draw->vs.clipdistance_output[index]; } /** * Return a pointer/handle for a driver/CSO rasterizer object which * disabled culling, stippling, unfilled tris, etc. * This is used by some pipeline stages (such as wide_point, aa_line * and aa_point) which convert points/lines into triangles. In those * cases we don't want to accidentally cull the triangles. * * \param scissor should the rasterizer state enable scissoring? * \param flatshade should the rasterizer state use flat shading? * \return rasterizer CSO handle */ void * draw_get_rasterizer_no_cull( struct draw_context *draw, boolean scissor, boolean flatshade ) { if (!draw->rasterizer_no_cull[scissor][flatshade]) { /* create now */ struct pipe_context *pipe = draw->pipe; struct pipe_rasterizer_state rast; memset(&rast, 0, sizeof(rast)); rast.scissor = scissor; rast.flatshade = flatshade; rast.front_ccw = 1; rast.half_pixel_center = draw->rasterizer->half_pixel_center; rast.bottom_edge_rule = draw->rasterizer->bottom_edge_rule; rast.clip_halfz = draw->rasterizer->clip_halfz; draw->rasterizer_no_cull[scissor][flatshade] = pipe->create_rasterizer_state(pipe, &rast); } return draw->rasterizer_no_cull[scissor][flatshade]; } void draw_set_mapped_so_targets(struct draw_context *draw, int num_targets, struct draw_so_target *targets[PIPE_MAX_SO_BUFFERS]) { int i; for (i = 0; i < num_targets; i++) draw->so.targets[i] = targets[i]; for (i = num_targets; i < PIPE_MAX_SO_BUFFERS; i++) draw->so.targets[i] = NULL; draw->so.num_targets = num_targets; } void draw_set_sampler_views(struct draw_context *draw, unsigned shader_stage, struct pipe_sampler_view **views, unsigned num) { unsigned i; debug_assert(shader_stage < PIPE_SHADER_TYPES); debug_assert(num <= PIPE_MAX_SHADER_SAMPLER_VIEWS); draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); for (i = 0; i < num; ++i) draw->sampler_views[shader_stage][i] = views[i]; for (i = num; i < PIPE_MAX_SHADER_SAMPLER_VIEWS; ++i) draw->sampler_views[shader_stage][i] = NULL; draw->num_sampler_views[shader_stage] = num; } void draw_set_samplers(struct draw_context *draw, unsigned shader_stage, struct pipe_sampler_state **samplers, unsigned num) { unsigned i; debug_assert(shader_stage < PIPE_SHADER_TYPES); debug_assert(num <= PIPE_MAX_SAMPLERS); draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE ); for (i = 0; i < num; ++i) draw->samplers[shader_stage][i] = samplers[i]; for (i = num; i < PIPE_MAX_SAMPLERS; ++i) draw->samplers[shader_stage][i] = NULL; draw->num_samplers[shader_stage] = num; #ifdef HAVE_LLVM if (draw->llvm) draw_llvm_set_sampler_state(draw, shader_stage); #endif } void draw_set_mapped_texture(struct draw_context *draw, unsigned shader_stage, unsigned sview_idx, uint32_t width, uint32_t height, uint32_t depth, uint32_t first_level, uint32_t last_level, const void *base_ptr, uint32_t row_stride[PIPE_MAX_TEXTURE_LEVELS], uint32_t img_stride[PIPE_MAX_TEXTURE_LEVELS], uint32_t mip_offsets[PIPE_MAX_TEXTURE_LEVELS]) { #ifdef HAVE_LLVM if (draw->llvm) draw_llvm_set_mapped_texture(draw, shader_stage, sview_idx, width, height, depth, first_level, last_level, base_ptr, row_stride, img_stride, mip_offsets); #endif } /** * XXX: Results for PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS because there are two * different ways of setting textures, and drivers typically only support one. */ int draw_get_shader_param_no_llvm(unsigned shader, enum pipe_shader_cap param) { switch(shader) { case PIPE_SHADER_VERTEX: case PIPE_SHADER_GEOMETRY: return tgsi_exec_get_shader_param(param); default: return 0; } } /** * XXX: Results for PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS because there are two * different ways of setting textures, and drivers typically only support one. */ int draw_get_shader_param(unsigned shader, enum pipe_shader_cap param) { #ifdef HAVE_LLVM if (draw_get_option_use_llvm()) { switch(shader) { case PIPE_SHADER_VERTEX: case PIPE_SHADER_GEOMETRY: return gallivm_get_shader_param(param); default: return 0; } } #endif return draw_get_shader_param_no_llvm(shader, param); } /** * Enables or disables collection of statistics. * * Draw module is capable of generating statistics for the vertex * processing pipeline. Collection of that data isn't free and so * it's disabled by default. The users of the module can enable * (or disable) this functionality through this function. * The actual data will be emitted through the VBUF interface, * the 'pipeline_statistics' callback to be exact. */ void draw_collect_pipeline_statistics(struct draw_context *draw, boolean enable) { draw->collect_statistics = enable; }