/**************************************************************************
*
- * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * Copyright 2007 VMware, Inc.
* All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* 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
+ * IN NO EVENT SHALL VMWARE 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 <keith@tungstengraphics.com>
+ * Keith Whitwell <keithw@vmware.com>
*/
#include "util/u_cpu_detect.h"
#include "util/u_inlines.h"
#include "util/u_helpers.h"
+#include "util/u_prim.h"
+#include "util/u_format.h"
#include "draw_context.h"
+#include "draw_pipe.h"
+#include "draw_prim_assembler.h"
#include "draw_vs.h"
#include "draw_gs.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;
+ return debug_get_bool_option("DRAW_USE_LLVM", TRUE);
+}
+#else
+boolean
+draw_get_option_use_llvm(void)
+{
+ return FALSE;
}
#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)
+draw_create_context(struct pipe_context *pipe, void *context,
+ boolean try_llvm)
{
struct draw_context *draw = CALLOC_STRUCT( draw_context );
- if (draw == NULL)
+ if (!draw)
goto err_out;
+ /* we need correct cpu caps for disabling denorms in draw_vbo() */
+ util_cpu_detect();
+
#if HAVE_LLVM
if (try_llvm && draw_get_option_use_llvm()) {
- draw->llvm = draw_llvm_create(draw);
- if (!draw->llvm)
- goto err_destroy;
+ draw->llvm = draw_llvm_create(draw, (LLVMContextRef)context);
}
#endif
if (!draw_init(draw))
goto err_destroy;
+ draw->ia = draw_prim_assembler_create(draw);
+ if (!draw->ia)
+ goto err_destroy;
+
return draw;
err_destroy:
struct draw_context *
draw_create(struct pipe_context *pipe)
{
- return draw_create_context(pipe, TRUE);
+ return draw_create_context(pipe, NULL, TRUE);
}
+#if HAVE_LLVM
+struct draw_context *
+draw_create_with_llvm_context(struct pipe_context *pipe,
+ void *context)
+{
+ return draw_create_context(pipe, context, TRUE);
+}
+#endif
+
/**
* 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);
+ return draw_create_context(pipe, NULL, FALSE);
}
draw->clip_z = TRUE;
draw->pt.user.planes = (float (*) [DRAW_TOTAL_CLIP_PLANES][4]) &(draw->plane[0]);
+ draw->pt.user.eltMax = ~0;
if (!draw_pipeline_init( draw ))
return FALSE;
draw->quads_always_flatshade_last = !draw->pipe->screen->get_param(
draw->pipe->screen, PIPE_CAP_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION);
+ draw->floating_point_depth = false;
+
return TRUE;
}
void draw_new_instance(struct draw_context *draw)
{
draw_geometry_shader_new_instance(draw->gs.geometry_shader);
+ draw_prim_assembler_new_instance(draw->ia);
}
}
}
- for (i = 0; i < draw->pt.nr_vertex_buffers; i++) {
- pipe_resource_reference(&draw->pt.vertex_buffer[i].buffer, NULL);
- }
+ for (i = 0; i < draw->pt.nr_vertex_buffers; i++)
+ pipe_vertex_buffer_unreference(&draw->pt.vertex_buffer[i]);
/* Not so fast -- we're just borrowing this at the moment.
*
draw->render->destroy( draw->render );
*/
+ draw_prim_assembler_destroy(draw->ia);
draw_pipeline_destroy( draw );
draw_pt_destroy( draw );
draw_vs_destroy( draw );
/**
- * Specify the Minimum Resolvable Depth factor for polygon offset.
+ * Specify the depth stencil format for the draw pipeline. This function
+ * determines 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.
+ * values between vertex shading and rasterization.
*/
-void draw_set_mrd(struct draw_context *draw, double mrd)
+void draw_set_zs_format(struct draw_context *draw, enum pipe_format format)
+{
+ const struct util_format_description *desc = util_format_description(format);
+
+ draw->floating_point_depth =
+ (util_get_depth_format_type(desc) == UTIL_FORMAT_TYPE_FLOAT);
+
+ draw->mrd = util_get_depth_format_mrd(desc);
+}
+
+
+static bool
+draw_is_vs_window_space(struct draw_context *draw)
{
- draw->mrd = mrd;
+ if (draw->vs.vertex_shader) {
+ struct tgsi_shader_info *info = &draw->vs.vertex_shader->info;
+
+ return info->properties[TGSI_PROPERTY_VS_WINDOW_SPACE_POSITION] != 0;
+ }
+ return false;
}
-static void update_clip_flags( struct draw_context *draw )
+void
+draw_update_clip_flags(struct draw_context *draw)
{
- draw->clip_xy = !draw->driver.bypass_clip_xy;
+ bool window_space = draw_is_vs_window_space(draw);
+
+ draw->clip_xy = !draw->driver.bypass_clip_xy && !window_space;
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->rasterizer && draw->rasterizer->depth_clip_near) &&
+ !window_space;
draw->clip_user = draw->rasterizer &&
- draw->rasterizer->clip_plane_enable != 0;
+ draw->rasterizer->clip_plane_enable != 0 &&
+ !window_space;
+ draw->guard_band_points_xy = draw->guard_band_xy ||
+ (draw->driver.bypass_clip_points &&
+ (draw->rasterizer &&
+ draw->rasterizer->point_tri_clip));
+}
+
+
+void
+draw_update_viewport_flags(struct draw_context *draw)
+{
+ bool window_space = draw_is_vs_window_space(draw);
+
+ draw->bypass_viewport = window_space || draw->identity_viewport;
}
+
/**
* Register new primitive rasterization/rendering state.
* This causes the drawing pipeline to be rebuilt.
draw->rasterizer = raster;
draw->rast_handle = rast_handle;
- update_clip_flags(draw);
+ draw_update_clip_flags(draw);
}
}
* 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.
+ * Setting bypass_clip_points to achieve d3d-style point clipping (the driver
+ * will need to do the "vp scissoring") _requires_ the driver to implement
+ * wide points / point sprites itself (points will still be clipped if rasterizer
+ * point_tri_clip isn't set). Only relevant if bypass_clip_xy isn't set.
*/
void draw_set_driver_clipping( struct draw_context *draw,
boolean bypass_clip_xy,
boolean bypass_clip_z,
- boolean guard_band_xy)
+ boolean guard_band_xy,
+ boolean bypass_clip_points)
{
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);
+ draw->driver.bypass_clip_points = bypass_clip_points;
+ draw_update_clip_flags(draw);
}
const struct pipe_viewport_state *viewport = vps;
draw_do_flush(draw, DRAW_FLUSH_PARAMETER_CHANGE);
- if (start_slot > PIPE_MAX_VIEWPORTS)
- return;
-
- if ((start_slot + num_viewports) > PIPE_MAX_VIEWPORTS)
- num_viewports = PIPE_MAX_VIEWPORTS - start_slot;
+ debug_assert(start_slot < PIPE_MAX_VIEWPORTS);
+ debug_assert((start_slot + num_viewports) <= PIPE_MAX_VIEWPORTS);
memcpy(draw->viewports + start_slot, vps,
sizeof(struct pipe_viewport_state) * num_viewports);
+
draw->identity_viewport = (num_viewports == 1) &&
(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);
+ viewport->translate[2] == 0.0f);
+ draw_update_viewport_flags(draw);
}
void
draw_set_mapped_constant_buffer(struct draw_context *draw,
- unsigned shader_type,
+ enum pipe_shader_type shader_type,
unsigned slot,
const void *buffer,
unsigned size )
}
}
+void
+draw_set_mapped_shader_buffer(struct draw_context *draw,
+ enum pipe_shader_type 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_SHADER_BUFFERS);
+
+ draw_do_flush(draw, DRAW_FLUSH_PARAMETER_CHANGE);
+
+ switch (shader_type) {
+ case PIPE_SHADER_VERTEX:
+ draw->pt.user.vs_ssbos[slot] = buffer;
+ draw->pt.user.vs_ssbos_size[slot] = size;
+ break;
+ case PIPE_SHADER_GEOMETRY:
+ draw->pt.user.gs_ssbos[slot] = buffer;
+ draw->pt.user.gs_ssbos_size[slot] = size;
+ break;
+ default:
+ assert(0 && "invalid shader type in draw_set_mapped_shader_buffer");
+ }
+}
/**
* Tells the draw module to draw points with triangles if their size
uint n;
slot = draw_find_shader_output(draw, semantic_name, semantic_index);
- if (slot > 0) {
+ 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));
+ assert(n < ARRAY_SIZE(draw->extra_shader_outputs.semantic_name));
draw->extra_shader_outputs.semantic_name[n] = semantic_name;
draw->extra_shader_outputs.semantic_index[n] = semantic_index;
}
}
+/**
+ * Prepare outputs slots from the draw module
+ *
+ * Certain parts of the draw module can emit additional
+ * outputs that can be quite useful to the backends, a good
+ * example of it is the process of decomposing primitives
+ * into wireframes (aka. lines) which normally would lose
+ * the face-side information, but using this method we can
+ * inject another shader output which passes the original
+ * face side information to the backend.
+ */
+void
+draw_prepare_shader_outputs(struct draw_context *draw)
+{
+ draw_remove_extra_vertex_attribs(draw);
+ draw_prim_assembler_prepare_outputs(draw->ia);
+ draw_unfilled_prepare_outputs(draw, draw->pipeline.unfilled);
+ if (draw->pipeline.aapoint)
+ draw_aapoint_prepare_outputs(draw, draw->pipeline.aapoint);
+ if (draw->pipeline.aaline)
+ draw_aaline_prepare_outputs(draw, draw->pipeline.aaline);
+}
/**
* Ask the draw module for the location/slot of the given vertex attribute in
* 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.
+ * -1 is returned if the attribute is not found since this is
+ * an undefined situation. Note, that zero is valid and can
+ * be used by any of the attributes, because position is not
+ * required to be attribute 0 or even at all present.
*/
int
draw_find_shader_output(const struct draw_context *draw,
}
}
- return 0;
+ return -1;
}
}
+/**
+ * Return total number of the vertex shader outputs. This function
+ * also counts any extra vertex output attributes that may
+ * be filled in by some draw stages (such as AA point, AA line,
+ * front face).
+ */
+uint
+draw_total_vs_outputs(const struct draw_context *draw)
+{
+ const struct tgsi_shader_info *info = &draw->vs.vertex_shader->info;
+
+ return info->num_outputs + draw->extra_shader_outputs.num;
+}
+
+/**
+ * Return total number of the geometry shader outputs. This function
+ * also counts any extra geometry output attributes that may
+ * be filled in by some draw stages (such as AA point, AA line, front
+ * face).
+ */
+uint
+draw_total_gs_outputs(const struct draw_context *draw)
+{
+ const struct tgsi_shader_info *info;
+
+ if (!draw->gs.geometry_shader)
+ return 0;
+
+ info = &draw->gs.geometry_shader->info;
+
+ return info->num_outputs + draw->extra_shader_outputs.num;
+}
+
+
/**
* Provide TGSI sampler objects for vertex/geometry shaders that use
* texture fetches. This state only needs to be set once per context.
*/
void
draw_texture_sampler(struct draw_context *draw,
- uint shader,
+ enum pipe_shader_type shader,
struct tgsi_sampler *sampler)
{
if (shader == PIPE_SHADER_VERTEX) {
}
}
+/**
+ * Provide TGSI image 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_image(struct draw_context *draw,
+ enum pipe_shader_type shader,
+ struct tgsi_image *image)
+{
+ if (shader == PIPE_SHADER_VERTEX) {
+ draw->vs.tgsi.image = image;
+ } else {
+ debug_assert(shader == PIPE_SHADER_GEOMETRY);
+ draw->gs.tgsi.image = image;
+ }
+}
+/**
+ * Provide TGSI buffer objects for vertex/geometry shaders that use
+ * load/store/atomic ops. This state only needs to be set once per context.
+ * This might only be used by software drivers for the time being.
+ */
+void
+draw_buffer(struct draw_context *draw,
+ enum pipe_shader_type shader,
+ struct tgsi_buffer *buffer)
+{
+ if (shader == PIPE_SHADER_VERTEX) {
+ draw->vs.tgsi.buffer = buffer;
+ } else {
+ debug_assert(shader == PIPE_SHADER_GEOMETRY);
+ draw->gs.tgsi.buffer = buffer;
+ }
+}
void draw_set_render( struct draw_context *draw,
/**
* 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,
{
if (draw->gs.geometry_shader)
return draw->gs.geometry_shader->viewport_index_output;
- return 0;
+ return draw->vs.vertex_shader->viewport_index_output;
}
/**
{
if (draw->gs.geometry_shader)
return draw->gs.geometry_shader->info.writes_viewport_index;
- return FALSE;
+ return draw->vs.vertex_shader->info.writes_viewport_index;
}
/**
* Return the index of the shader output which will contain the
- * vertex position.
+ * clip vertex position.
+ * Note we don't support clipvertex output in the gs. For clipping
+ * to work correctly hence we return ordinary position output instead.
*/
uint
draw_current_shader_clipvertex_output(const struct draw_context *draw)
{
+ if (draw->gs.geometry_shader)
+ return draw->gs.position_output;
return draw->vs.clipvertex_output;
}
uint
-draw_current_shader_clipdistance_output(const struct draw_context *draw, int index)
+draw_current_shader_ccdistance_output(const struct draw_context *draw, int index)
{
- return draw->vs.clipdistance_output[index];
+ debug_assert(index < PIPE_MAX_CLIP_OR_CULL_DISTANCE_ELEMENT_COUNT);
+ if (draw->gs.geometry_shader)
+ return draw->gs.geometry_shader->ccdistance_output[index];
+ return draw->vs.ccdistance_output[index];
+}
+
+
+uint
+draw_current_shader_num_written_clipdistances(const struct draw_context *draw)
+{
+ if (draw->gs.geometry_shader)
+ return draw->gs.geometry_shader->info.num_written_clipdistance;
+ return draw->vs.vertex_shader->info.num_written_clipdistance;
+}
+
+uint
+draw_current_shader_num_written_culldistances(const struct draw_context *draw)
+{
+ if (draw->gs.geometry_shader)
+ return draw->gs.geometry_shader->info.num_written_culldistance;
+ return draw->vs.vertex_shader->info.num_written_culldistance;
}
/**
{
int i;
+ draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE );
+
for (i = 0; i < num_targets; i++)
draw->so.targets[i] = targets[i];
for (i = num_targets; i < PIPE_MAX_SO_BUFFERS; i++)
void
draw_set_sampler_views(struct draw_context *draw,
- unsigned shader_stage,
+ enum pipe_shader_type shader_stage,
struct pipe_sampler_view **views,
unsigned num)
{
for (i = 0; i < num; ++i)
draw->sampler_views[shader_stage][i] = views[i];
- for (i = num; i < PIPE_MAX_SHADER_SAMPLER_VIEWS; ++i)
+ for (i = num; i < draw->num_sampler_views[shader_stage]; ++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,
+ enum pipe_shader_type shader_stage,
struct pipe_sampler_state **samplers,
unsigned num)
{
#endif
}
+void
+draw_set_images(struct draw_context *draw,
+ enum pipe_shader_type shader_stage,
+ struct pipe_image_view *views,
+ unsigned num)
+{
+ unsigned i;
+
+ debug_assert(shader_stage < PIPE_SHADER_TYPES);
+ debug_assert(num <= PIPE_MAX_SHADER_IMAGES);
+
+ draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE );
+
+ for (i = 0; i < num; ++i)
+ draw->images[shader_stage][i] = &views[i];
+ for (i = num; i < draw->num_sampler_views[shader_stage]; ++i)
+ draw->images[shader_stage][i] = NULL;
+
+ draw->num_images[shader_stage] = num;
+}
+
void
draw_set_mapped_texture(struct draw_context *draw,
- unsigned shader_stage,
+ enum pipe_shader_type shader_stage,
unsigned sview_idx,
uint32_t width, uint32_t height, uint32_t depth,
uint32_t first_level, uint32_t last_level,
#endif
}
+void
+draw_set_mapped_image(struct draw_context *draw,
+ enum pipe_shader_type shader_stage,
+ unsigned idx,
+ uint32_t width, uint32_t height, uint32_t depth,
+ const void *base_ptr,
+ uint32_t row_stride,
+ uint32_t img_stride)
+{
+#ifdef HAVE_LLVM
+ if (draw->llvm)
+ draw_llvm_set_mapped_image(draw,
+ shader_stage,
+ idx,
+ width, height, depth,
+ base_ptr,
+ row_stride, img_stride);
+#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)
+draw_get_shader_param_no_llvm(enum pipe_shader_type shader,
+ enum pipe_shader_cap param)
{
switch(shader) {
case PIPE_SHADER_VERTEX:
/**
* XXX: Results for PIPE_SHADER_CAP_MAX_TEXTURE_SAMPLERS because there are two
* different ways of setting textures, and drivers typically only support one.
+ * Drivers requesting a draw context explicitly without llvm must call
+ * draw_get_shader_param_no_llvm instead.
*/
int
-draw_get_shader_param(unsigned shader, enum pipe_shader_cap param)
+draw_get_shader_param(enum pipe_shader_type shader, enum pipe_shader_cap param)
{
#ifdef HAVE_LLVM
{
draw->collect_statistics = enable;
}
+
+/**
+ * Computes clipper invocation statistics.
+ *
+ * Figures out how many primitives would have been
+ * sent to the clipper given the specified
+ * prim info data.
+ */
+void
+draw_stats_clipper_primitives(struct draw_context *draw,
+ const struct draw_prim_info *prim_info)
+{
+ if (draw->collect_statistics) {
+ unsigned i;
+ for (i = 0; i < prim_info->primitive_count; i++) {
+ draw->statistics.c_invocations +=
+ u_decomposed_prims_for_vertices(prim_info->prim,
+ prim_info->primitive_lengths[i]);
+ }
+ }
+}
+
+
+/**
+ * Returns true if the draw module will inject the frontface
+ * info into the outputs.
+ *
+ * Given the specified primitive and rasterizer state
+ * the function will figure out if the draw module
+ * will inject the front-face information into shader
+ * outputs. This is done to preserve the front-facing
+ * info when decomposing primitives into wireframes.
+ */
+boolean
+draw_will_inject_frontface(const struct draw_context *draw)
+{
+ unsigned reduced_prim = u_reduced_prim(draw->pt.prim);
+ const struct pipe_rasterizer_state *rast = draw->rasterizer;
+
+ if (reduced_prim != PIPE_PRIM_TRIANGLES) {
+ return FALSE;
+ }
+
+ return (rast &&
+ (rast->fill_front != PIPE_POLYGON_MODE_FILL ||
+ rast->fill_back != PIPE_POLYGON_MODE_FILL));
+}
projects / mesa.git / blobdiff