On Thu, Oct 20, 2016 at 8:23 PM, Timothy Arceri <
timothy.arc...@collabora.com> wrote:
> Here brw_setup_vue_interpolation() is rewritten not to use the
> InterpQualifier
> array in gl_fragment_program which will allow us to remove it.
>
> This change also makes the code which is only used by gen4/5 more self
> contained
> as it now has its own gen5_fragment_program struct rather than storing the
> map
> in brw_context. This means the interpolation map will only get processed
> once
> and will get stored in the in memory cache rather than being processed
> everytime
> the fs changes.
>
> Also by calling this from the fs compile code rather than from the upload
> code
> and using the interpolation assigned there we can get rid of the
> BRW_NEW_INTERPOLATION_MAP flag.
>
> It might not seem ideal to add a gen5_fragment_program struct however by
> the end
> of this series we will have gotten rid of all the
> brw_{shader_stage}_program
> structs and replaced them with a generic brw_program struct so there will
> only
> be two program structs which is better than what we have now.
>
> V2: Don't remove BRW_NEW_INTERPOLATION_MAP from dirty_bit_map until the
> following
> patch to fix build error.
> ---
> src/intel/blorp/blorp.c | 6 +-
> src/intel/vulkan/anv_pipeline.c | 2 +-
> src/mesa/drivers/dri/i965/brw_clip.c | 19 ++---
> src/mesa/drivers/dri/i965/brw_clip.h | 7 +-
> src/mesa/drivers/dri/i965/brw_clip_line.c | 2 +-
> src/mesa/drivers/dri/i965/brw_clip_tri.c | 2 +-
> src/mesa/drivers/dri/i965/brw_clip_unfilled.c | 2 +-
> src/mesa/drivers/dri/i965/brw_clip_util.c | 10 +--
> src/mesa/drivers/dri/i965/brw_compiler.h | 8 ++-
> src/mesa/drivers/dri/i965/brw_context.h | 45 ++++--------
> src/mesa/drivers/dri/i965/brw_fs.cpp | 4 +-
> src/mesa/drivers/dri/i965/brw_interpolation_map.c | 85
> +++++++++++------------
> src/mesa/drivers/dri/i965/brw_nir.c | 7 +-
> src/mesa/drivers/dri/i965/brw_nir.h | 3 +-
> src/mesa/drivers/dri/i965/brw_program.c | 10 ++-
> src/mesa/drivers/dri/i965/brw_sf.c | 14 ++--
> src/mesa/drivers/dri/i965/brw_sf.h | 4 +-
> src/mesa/drivers/dri/i965/brw_sf_emit.c | 12 ++--
> src/mesa/drivers/dri/i965/brw_state.h | 3 -
> src/mesa/drivers/dri/i965/brw_state_upload.c | 5 +-
> src/mesa/drivers/dri/i965/brw_wm.c | 18 +++--
> src/mesa/drivers/dri/i965/brw_wm.h | 3 +-
> 22 files changed, 142 insertions(+), 129 deletions(-)
>
> diff --git a/src/intel/blorp/blorp.c b/src/intel/blorp/blorp.c
> index 5209ee2..8905cfa 100644
> --- a/src/intel/blorp/blorp.c
> +++ b/src/intel/blorp/blorp.c
> @@ -211,9 +211,9 @@ brw_blorp_compile_nir_shader(struct blorp_context
> *blorp, struct nir_shader *nir
> nir_lower_io(nir, nir_var_uniform, nir_uniform_type_size, 0);
>
> const unsigned *program =
> - brw_compile_fs(compiler, blorp->driver_ctx, mem_ctx,
> - wm_key, &wm_prog_data, nir,
> - NULL, -1, -1, false, use_repclear, program_size,
> NULL);
> + brw_compile_fs(compiler, blorp->driver_ctx, mem_ctx, wm_key,
> + &wm_prog_data, nir, NULL, -1, -1, false,
> use_repclear,
> + NULL, program_size, NULL);
>
> /* Copy the relavent bits of wm_prog_data over into the blorp prog
> data */
> prog_data->dispatch_8 = wm_prog_data.dispatch_8;
> diff --git a/src/intel/vulkan/anv_pipeline.c b/src/intel/vulkan/anv_
> pipeline.c
> index 72f0643..0aac711 100644
> --- a/src/intel/vulkan/anv_pipeline.c
> +++ b/src/intel/vulkan/anv_pipeline.c
> @@ -678,7 +678,7 @@ anv_pipeline_compile_fs(struct anv_pipeline *pipeline,
> unsigned code_size;
> const unsigned *shader_code =
> brw_compile_fs(compiler, NULL, mem_ctx, &key, &prog_data, nir,
> - NULL, -1, -1, true, false, &code_size, NULL);
> + NULL, -1, -1, true, false, NULL, &code_size,
> NULL);
> if (shader_code == NULL) {
> ralloc_free(mem_ctx);
> return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
> diff --git a/src/mesa/drivers/dri/i965/brw_clip.c
> b/src/mesa/drivers/dri/i965/brw_clip.c
> index 1134fa4..06650b6 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip.c
> +++ b/src/mesa/drivers/dri/i965/brw_clip.c
> @@ -68,11 +68,6 @@ static void compile_clip_prog( struct brw_context *brw,
> c.key = *key;
> c.vue_map = brw->vue_map_geom_out;
>
> - c.has_flat_shading =
> - brw_any_flat_varyings(&key->interpolation_mode);
> - c.has_noperspective_shading =
> - brw_any_noperspective_varyings(&key->interpolation_mode);
> -
> /* nr_regs is the number of registers filled by reading data from the
> VUE.
> * This program accesses the entire VUE, so nr_regs needs to be the
> size of
> * the VUE (measured in pairs, since two slots are stored in each
> @@ -144,7 +139,7 @@ brw_upload_clip_prog(struct brw_context *brw)
> _NEW_POLYGON |
> _NEW_TRANSFORM,
> BRW_NEW_BLORP |
> - BRW_NEW_INTERPOLATION_MAP |
> + BRW_NEW_FRAGMENT_PROGRAM |
> BRW_NEW_REDUCED_PRIMITIVE |
> BRW_NEW_VUE_MAP_GEOM_OUT))
> return;
> @@ -154,8 +149,16 @@ brw_upload_clip_prog(struct brw_context *brw)
> /* Populate the key:
> */
>
> - /* BRW_NEW_INTERPOLATION_MAP */
> - key.interpolation_mode = brw->interpolation_mode;
> + const struct gl_fragment_program *fprog = brw->fragment_program;
> + if (fprog) {
> + assert(brw->gen < 6);
> + struct gen5_fragment_program *p = (struct gen5_fragment_program *)
> fprog;
> +
> + /* BRW_NEW_FRAGMENT_PROGRAM */
> + key.contains_flat_varying = p->contains_flat_varying;
> + key.contains_noperspective_varying = p->contains_noperspective_
> varying;
> + key.interp_mode = p->interp_mode;
> + }
>
> /* BRW_NEW_REDUCED_PRIMITIVE */
> key.primitive = brw->reduced_primitive;
> diff --git a/src/mesa/drivers/dri/i965/brw_clip.h
> b/src/mesa/drivers/dri/i965/brw_clip.h
> index 54c7682..355ae64 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip.h
> +++ b/src/mesa/drivers/dri/i965/brw_clip.h
> @@ -47,7 +47,9 @@
> */
> struct brw_clip_prog_key {
> GLbitfield64 attrs;
> - struct interpolation_mode_map interpolation_mode;
> + bool contains_flat_varying;
> + bool contains_noperspective_varying;
> + unsigned char *interp_mode;
> GLuint primitive:4;
> GLuint nr_userclip:4;
> GLuint pv_first:1;
> @@ -128,9 +130,6 @@ struct brw_clip_compile {
> bool need_direction;
>
> struct brw_vue_map vue_map;
> -
> - bool has_flat_shading;
> - bool has_noperspective_shading;
> };
>
> /**
> diff --git a/src/mesa/drivers/dri/i965/brw_clip_line.c
> b/src/mesa/drivers/dri/i965/brw_clip_line.c
> index 2c7f85a..f9cabd2 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip_line.c
> +++ b/src/mesa/drivers/dri/i965/brw_clip_line.c
> @@ -302,7 +302,7 @@ void brw_emit_line_clip( struct brw_clip_compile *c )
> brw_clip_line_alloc_regs(c);
> brw_clip_init_ff_sync(c);
>
> - if (c->has_flat_shading) {
> + if (c->key.contains_flat_varying) {
> if (c->key.pv_first)
> brw_clip_copy_flatshaded_attributes(c, 1, 0);
> else
> diff --git a/src/mesa/drivers/dri/i965/brw_clip_tri.c
> b/src/mesa/drivers/dri/i965/brw_clip_tri.c
> index 4caa8f5..52402e3 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip_tri.c
> +++ b/src/mesa/drivers/dri/i965/brw_clip_tri.c
> @@ -650,7 +650,7 @@ void brw_emit_tri_clip( struct brw_clip_compile *c )
> * flatshading, need to apply the flatshade here because we don't
> * respect the PV when converting to trifan for emit:
> */
> - if (c->has_flat_shading)
> + if (c->key.contains_flat_varying)
> brw_clip_tri_flat_shade(c);
>
> if ((c->key.clip_mode == BRW_CLIPMODE_NORMAL) ||
> diff --git a/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
> b/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
> index d333d10..7c06857 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
> +++ b/src/mesa/drivers/dri/i965/brw_clip_unfilled.c
> @@ -519,7 +519,7 @@ void brw_emit_unfilled_clip( struct brw_clip_compile
> *c )
>
> /* Need to do this whether we clip or not:
> */
> - if (c->has_flat_shading)
> + if (c->key.contains_flat_varying)
> brw_clip_tri_flat_shade(c);
>
> brw_clip_init_clipmask(c);
> diff --git a/src/mesa/drivers/dri/i965/brw_clip_util.c
> b/src/mesa/drivers/dri/i965/brw_clip_util.c
> index f5d0a5a..4f1b0fd 100644
> --- a/src/mesa/drivers/dri/i965/brw_clip_util.c
> +++ b/src/mesa/drivers/dri/i965/brw_clip_util.c
> @@ -157,7 +157,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile
> *c,
> */
>
> /* Take a copy of the v0 NDC coordinates, in case dest == v0. */
> - if (c->has_noperspective_shading) {
> + if (c->key.contains_noperspective_varying) {
> GLuint offset = brw_varying_to_offset(&c->vue_map,
> BRW_VARYING_SLOT_NDC);
> v0_ndc_copy = get_tmp(c);
> @@ -183,7 +183,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile
> *c,
> /* If we have noperspective attributes,
> * we need to compute the screen-space t
> */
> - if (c->has_noperspective_shading) {
> + if (c->key.contains_noperspective_varying) {
> GLuint delta = brw_varying_to_offset(&c->vue_map,
> BRW_VARYING_SLOT_NDC);
> struct brw_reg tmp = get_tmp(c);
> @@ -272,7 +272,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile
> *c,
> * Unless the attribute is flat shaded -- in which case just copy
> * from one of the sources (doesn't matter which; already copied
> from pv)
> */
> - GLuint interp = c->key.interpolation_mode.mode[slot];
> + GLuint interp = c->key.interp_mode[slot];
>
> if (interp != INTERP_MODE_FLAT) {
> struct brw_reg tmp = get_tmp(c);
> @@ -310,7 +310,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile
> *c,
> brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(0));
> }
>
> - if (c->has_noperspective_shading)
> + if (c->key.contains_noperspective_varying)
> release_tmp(c, t_nopersp);
> }
>
> @@ -406,7 +406,7 @@ void brw_clip_copy_flatshaded_attributes( struct
> brw_clip_compile *c,
> struct brw_codegen *p = &c->func;
>
> for (int i = 0; i < c->vue_map.num_slots; i++) {
> - if (c->key.interpolation_mode.mode[i] == INTERP_MODE_FLAT) {
> + if (c->key.interp_mode[i] == INTERP_MODE_FLAT) {
> brw_MOV(p,
> byte_offset(c->reg.vertex[to],
> brw_vue_slot_to_offset(i)),
> byte_offset(c->reg.vertex[from],
> brw_vue_slot_to_offset(i)));
> diff --git a/src/mesa/drivers/dri/i965/brw_compiler.h
> b/src/mesa/drivers/dri/i965/brw_compiler.h
> index 447d05b..772c11e 100644
> --- a/src/mesa/drivers/dri/i965/brw_compiler.h
> +++ b/src/mesa/drivers/dri/i965/brw_compiler.h
> @@ -575,6 +575,12 @@ void brw_compute_tess_vue_map(struct brw_vue_map
> *const vue_map,
> const GLbitfield64 slots_valid,
> const GLbitfield is_patch);
>
> +/* brw_interpolation_map.c */
> +void brw_setup_vue_interpolation(struct brw_vue_map *vue_map,
> + struct nir_shader *nir,
> + struct gl_program *prog,
> + const struct gen_device_info *devinfo);
> +
> enum shader_dispatch_mode {
> DISPATCH_MODE_4X1_SINGLE = 0,
> DISPATCH_MODE_4X2_DUAL_INSTANCE = 1,
> @@ -834,7 +840,7 @@ brw_compile_fs(const struct brw_compiler *compiler,
> void *log_data,
> int shader_time_index8,
> int shader_time_index16,
> bool allow_spilling,
> - bool use_rep_send,
> + bool use_rep_send, struct brw_vue_map *vue_map,
> unsigned *final_assembly_size,
> char **error_str);
>
> diff --git a/src/mesa/drivers/dri/i965/brw_context.h
> b/src/mesa/drivers/dri/i965/brw_context.h
> index e190d53..ba8b254 100644
> --- a/src/mesa/drivers/dri/i965/brw_context.h
> +++ b/src/mesa/drivers/dri/i965/brw_context.h
> @@ -363,6 +363,20 @@ struct brw_fragment_program {
> };
>
>
> +struct gen5_fragment_program {
> + struct brw_fragment_program base;
> +
> + bool contains_flat_varying;
> + bool contains_noperspective_varying;
> +
> + /*
> + * Mapping of varying slots to interpolation modes.
> + * Used Gen4/5 by the clip|sf|wm stages.
> + */
> + unsigned char interp_mode[BRW_VARYING_SLOT_COUNT];
> +};
> +
> +
> /** Subclass of Mesa compute program */
> struct brw_compute_program {
> struct gl_program program;
> @@ -385,32 +399,6 @@ struct brw_shader {
> ~VARYING_BIT_POS & ~VARYING_BIT_FACE)
>
>
> -/*
> - * Mapping of VUE map slots to interpolation modes.
> - */
> -struct interpolation_mode_map {
> - unsigned char mode[BRW_VARYING_SLOT_COUNT];
> -};
> -
> -static inline bool brw_any_flat_varyings(struct interpolation_mode_map
> *map)
> -{
> - for (int i = 0; i < BRW_VARYING_SLOT_COUNT; i++)
> - if (map->mode[i] == INTERP_MODE_FLAT)
> - return true;
> -
> - return false;
> -}
> -
> -static inline bool brw_any_noperspective_varyings(struct
> interpolation_mode_map *map)
> -{
> - for (int i = 0; i < BRW_VARYING_SLOT_COUNT; i++)
> - if (map->mode[i] == INTERP_MODE_NOPERSPECTIVE)
> - return true;
> -
> - return false;
> -}
> -
> -
> struct brw_sf_prog_data {
> GLuint urb_read_length;
> GLuint total_grf;
> @@ -1273,11 +1261,6 @@ struct brw_context
> uint32_t render_target_format[MESA_FORMAT_COUNT];
> bool format_supported_as_render_target[MESA_FORMAT_COUNT];
>
> - /* Interpolation modes, one byte per vue slot.
> - * Used Gen4/5 by the clip|sf|wm stages. Ignored on Gen6+.
> - */
> - struct interpolation_mode_map interpolation_mode;
> -
> /* PrimitiveRestart */
> struct {
> bool in_progress;
> diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp
> b/src/mesa/drivers/dri/i965/brw_fs.cpp
> index 1a22fb4..3e0d057 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
> @@ -6422,14 +6422,14 @@ brw_compile_fs(const struct brw_compiler
> *compiler, void *log_data,
> struct gl_program *prog,
> int shader_time_index8, int shader_time_index16,
> bool allow_spilling,
> - bool use_rep_send,
> + bool use_rep_send, struct brw_vue_map *vue_map,
> unsigned *final_assembly_size,
> char **error_str)
> {
> nir_shader *shader = nir_shader_clone(mem_ctx, src_shader);
> shader = brw_nir_apply_sampler_key(shader, compiler->devinfo,
> &key->tex,
> true);
> - brw_nir_lower_fs_inputs(shader, compiler->devinfo, key);
> + brw_nir_lower_fs_inputs(shader, vue_map, prog, compiler->devinfo,
> key);
> brw_nir_lower_fs_outputs(shader);
> if (!key->multisample_fbo)
> NIR_PASS_V(shader, demote_sample_qualifiers);
> diff --git a/src/mesa/drivers/dri/i965/brw_interpolation_map.c
> b/src/mesa/drivers/dri/i965/brw_interpolation_map.c
> index 097987b..cb4deb5 100644
> --- a/src/mesa/drivers/dri/i965/brw_interpolation_map.c
> +++ b/src/mesa/drivers/dri/i965/brw_interpolation_map.c
> @@ -21,7 +21,8 @@
> * IN THE SOFTWARE.
> */
>
> -#include "brw_state.h"
> +#include "brw_compiler.h"
> +#include "brw_context.h"
> #include "compiler/nir/nir.h"
>
> static char const *get_qual_name(int mode)
> @@ -38,58 +39,52 @@ static char const *get_qual_name(int mode)
>
> /* Set up interpolation modes for every element in the VUE */
> void
> -brw_setup_vue_interpolation(struct brw_context *brw)
> +brw_setup_vue_interpolation(struct brw_vue_map *vue_map, nir_shader *nir,
> + struct gl_program *prog,
> + const struct gen_device_info *devinfo)
> {
> - const struct gl_fragment_program *fprog = brw->fragment_program;
> - struct brw_vue_map *vue_map = &brw->vue_map_geom_out;
> + struct gen5_fragment_program *fprog = (struct gen5_fragment_program *)
> prog;
>
Why are we calling this gen5_fragment_program and not
gen4_fragment_program? Generally, we use the lowest gen that applies.
> - if (!brw_state_dirty(brw,
> - _NEW_LIGHT,
> - BRW_NEW_BLORP |
> - BRW_NEW_FRAGMENT_PROGRAM |
> - BRW_NEW_VUE_MAP_GEOM_OUT))
> - return;
> -
> - memset(&brw->interpolation_mode, INTERP_MODE_NONE,
> sizeof(brw->interpolation_mode));
> -
> - brw->ctx.NewDriverState |= BRW_NEW_INTERPOLATION_MAP;
> + memset(fprog->interp_mode, INTERP_MODE_NONE,
> sizeof(fprog->interp_mode));
>
Please, no. This does not do what it looks like it does. Most enums
(unless declared as packed) aren't compiled down to 8-bit types so the only
reason why this works is because INTERP_MODE_NONE happens to be 0. If it
was anything else, this line wouldn't work. Memset to 0 with a comment or
use a for loop; don't memset to an enum value.
>
> - if (!fprog)
> + if (!vue_map)
> return;
>
> - for (int i = 0; i < vue_map->num_slots; i++) {
> - int varying = vue_map->slot_to_varying[i];
> - if (varying == -1)
> - continue;
> + /* HPOS always wants noperspective. setting it up here allows
> + * us to not need special handling in the SF program.
> + */
> + unsigned pos_slot = vue_map->varying_to_slot[VARYING_SLOT_POS];
> + if (pos_slot != -1) {;
> + fprog->interp_mode[pos_slot] = INTERP_MODE_NOPERSPECTIVE;
> + fprog->contains_noperspective_varying = true;
> + }
>
> - /* HPOS always wants noperspective. setting it up here allows
> - * us to not need special handling in the SF program. */
> - if (varying == VARYING_SLOT_POS) {
> - brw->interpolation_mode.mode[i] = INTERP_MODE_NOPERSPECTIVE;
> - continue;
> + foreach_list_typed(nir_variable, var, node, &nir->inputs) {
> + unsigned location = var->data.location;
> + unsigned slot_count = glsl_count_attribute_slots(var->type, false);
> + unsigned remap_count = 1;
> + if (location == VARYING_SLOT_COL0 || location == VARYING_SLOT_COL1)
> {
> + remap_count = 2;
> }
>
> - int frag_attrib = varying;
> - if (varying == VARYING_SLOT_BFC0 || varying == VARYING_SLOT_BFC1)
> - frag_attrib = varying - VARYING_SLOT_BFC0 + VARYING_SLOT_COL0;
> -
> - if (!(fprog->Base.nir->info->inputs_read &
> BITFIELD64_BIT(frag_attrib)))
> - continue;
> -
> - enum glsl_interp_mode mode = fprog->InterpQualifier[frag_attrib];
> -
> - /* If the mode is not specified, the default varies: Color values
> - * follow GL_SHADE_MODEL; everything else is smooth.
> - */
> - if (mode == INTERP_MODE_NONE) {
> - if (frag_attrib == VARYING_SLOT_COL0 || frag_attrib ==
> VARYING_SLOT_COL1)
> - mode = brw->ctx.Light.ShadeModel == GL_FLAT
> - ? INTERP_MODE_FLAT : INTERP_MODE_SMOOTH;
> - else
> - mode = INTERP_MODE_SMOOTH;
>
Uh... Where did this logic go? Oh, it's already in nir_lower_fs_inputs...
> + for (unsigned j = 0; j < remap_count; j++) {
>
This remap thing is really confusing. Maybe it would be better to have a
helper
gen5_fragment_prog_set_interp_modes(prog, vue_map, location, slot_count,
interp_mode)
And then, you can explicitly call it twice for colors and once for other
things. I believe your logic here is correct, it's just horribly confusing.
> + if (j == 1)
> + location = location + VARYING_SLOT_BFC0 - VARYING_SLOT_COL0;
> +
> + for (unsigned k = 0; k < slot_count; k++) {
> + unsigned slot = vue_map->varying_to_slot[location + k];
> + if (slot != -1 && fprog->interp_mode[slot] ==
> INTERP_MODE_NONE) {
> + fprog->interp_mode[slot] = var->data.interpolation;
> +
> + if (fprog->interp_mode[slot] == INTERP_MODE_FLAT) {
> + fprog->contains_flat_varying = true;
> + } else if (fprog->interp_mode[slot] ==
> + INTERP_MODE_NOPERSPECTIVE) {
> + fprog->contains_noperspective_varying = true;
> + }
> + }
> + }
> }
> -
> - brw->interpolation_mode.mode[i] = mode;
> }
>
> if (unlikely(INTEL_DEBUG & DEBUG_VUE)) {
> @@ -103,7 +98,7 @@ brw_setup_vue_interpolation(struct brw_context *brw)
>
> fprintf(stderr, "%d: %d %s ofs %d\n",
> i, varying,
> - get_qual_name(brw->interpolation_mode.mode[i]),
> + get_qual_name(fprog->interp_mode[i]),
> brw_vue_slot_to_offset(i));
> }
> }
> diff --git a/src/mesa/drivers/dri/i965/brw_nir.c
> b/src/mesa/drivers/dri/i965/brw_nir.c
> index 3d19691..e6566ba 100644
> --- a/src/mesa/drivers/dri/i965/brw_nir.c
> +++ b/src/mesa/drivers/dri/i965/brw_nir.c
> @@ -280,7 +280,8 @@ brw_nir_lower_tes_inputs(nir_shader *nir, const
> struct brw_vue_map *vue_map)
> }
>
> void
> -brw_nir_lower_fs_inputs(nir_shader *nir,
> +brw_nir_lower_fs_inputs(nir_shader *nir, struct brw_vue_map *vue_map,
> + struct gl_program *prog,
> const struct gen_device_info *devinfo,
> const struct brw_wm_prog_key *key)
> {
> @@ -311,6 +312,10 @@ brw_nir_lower_fs_inputs(nir_shader *nir,
> }
> }
>
> + if (devinfo->gen < 6) {
>
This is a lie. It really needs to be (devinfo->gen < 6 && prog). Or, you
could assert prog. Something to indicate that prog can't be null here.
> + brw_setup_vue_interpolation(vue_map, nir, prog, devinfo);
> + }
> +
> nir_lower_io_options lower_io_options = 0;
> if (key->persample_interp)
> lower_io_options |= nir_lower_io_force_sample_interpolation;
> diff --git a/src/mesa/drivers/dri/i965/brw_nir.h
> b/src/mesa/drivers/dri/i965/brw_nir.h
> index aef5c53..f989026 100644
> --- a/src/mesa/drivers/dri/i965/brw_nir.h
> +++ b/src/mesa/drivers/dri/i965/brw_nir.h
> @@ -105,7 +105,8 @@ void brw_nir_lower_vs_inputs(nir_shader *nir,
> void brw_nir_lower_vue_inputs(nir_shader *nir, bool is_scalar,
> const struct brw_vue_map *vue_map);
> void brw_nir_lower_tes_inputs(nir_shader *nir, const struct brw_vue_map
> *vue);
> -void brw_nir_lower_fs_inputs(nir_shader *nir,
> +void brw_nir_lower_fs_inputs(nir_shader *nir, struct brw_vue_map
> *vue_map,
> + struct gl_program *prog,
> const struct gen_device_info *devinfo,
> const struct brw_wm_prog_key *key);
> void brw_nir_lower_vue_outputs(nir_shader *nir, bool is_scalar);
> diff --git a/src/mesa/drivers/dri/i965/brw_program.c
> b/src/mesa/drivers/dri/i965/brw_program.c
> index 7516baf..957c5b5 100644
> --- a/src/mesa/drivers/dri/i965/brw_program.c
> +++ b/src/mesa/drivers/dri/i965/brw_program.c
> @@ -142,7 +142,15 @@ static struct gl_program *brwNewProgram( struct
> gl_context *ctx,
> }
>
> case GL_FRAGMENT_PROGRAM_ARB: {
> - struct brw_fragment_program *prog = CALLOC_STRUCT(brw_fragment_
> program);
> + struct brw_fragment_program *prog;
> + if (brw->gen < 6) {
> + struct gen5_fragment_program *g5_prog =
> + CALLOC_STRUCT(gen5_fragment_program);
> + prog = &g5_prog->base;
> + } else {
> + prog = CALLOC_STRUCT(brw_fragment_program);
> + }
> +
> if (prog) {
> prog->id = get_new_program_id(brw->screen);
>
> diff --git a/src/mesa/drivers/dri/i965/brw_sf.c
> b/src/mesa/drivers/dri/i965/brw_sf.c
> index 094260e..308b056 100644
> --- a/src/mesa/drivers/dri/i965/brw_sf.c
> +++ b/src/mesa/drivers/dri/i965/brw_sf.c
> @@ -79,7 +79,6 @@ static void compile_sf_prog( struct brw_context *brw,
>
> c.prog_data.urb_read_length = c.nr_attr_regs;
> c.prog_data.urb_entry_size = c.nr_setup_regs * 2;
> - c.has_flat_shading = brw_any_flat_varyings(&key->interpolation_mode);
>
> /* Which primitive? Or all three?
> */
> @@ -148,7 +147,7 @@ brw_upload_sf_prog(struct brw_context *brw)
> _NEW_PROGRAM |
> _NEW_TRANSFORM,
> BRW_NEW_BLORP |
> - BRW_NEW_INTERPOLATION_MAP |
> + BRW_NEW_FRAGMENT_PROGRAM |
> BRW_NEW_REDUCED_PRIMITIVE |
> BRW_NEW_VUE_MAP_GEOM_OUT))
> return;
> @@ -204,8 +203,15 @@ brw_upload_sf_prog(struct brw_context *brw)
> if ((ctx->Point.SpriteOrigin == GL_LOWER_LEFT) != render_to_fbo)
> key.sprite_origin_lower_left = true;
>
> - /* BRW_NEW_INTERPOLATION_MAP */
> - key.interpolation_mode = brw->interpolation_mode;
> + const struct gl_fragment_program *fprog = brw->fragment_program;
> + if (fprog) {
> + assert(brw->gen < 6);
> + struct gen5_fragment_program *p = (struct gen5_fragment_program *)
> fprog;
> +
> + /* BRW_NEW_FRAGMENT_PROGRAM */
> + key.contains_flat_varying = p->contains_flat_varying;
> + key.interp_mode = p->interp_mode;
> + }
>
> /* _NEW_LIGHT | _NEW_PROGRAM */
> key.do_twoside_color = ((ctx->Light.Enabled &&
> ctx->Light.Model.TwoSide) ||
> diff --git a/src/mesa/drivers/dri/i965/brw_sf.h
> b/src/mesa/drivers/dri/i965/brw_sf.h
> index 15102ac..ce4b2e3 100644
> --- a/src/mesa/drivers/dri/i965/brw_sf.h
> +++ b/src/mesa/drivers/dri/i965/brw_sf.h
> @@ -46,7 +46,8 @@
>
> struct brw_sf_prog_key {
> GLbitfield64 attrs;
> - struct interpolation_mode_map interpolation_mode;
> + bool contains_flat_varying;
> + unsigned char *interp_mode;
> uint8_t point_sprite_coord_replace;
> GLuint primitive:2;
> GLuint do_twoside_color:1;
> @@ -98,7 +99,6 @@ struct brw_sf_compile {
> unsigned flag_value;
>
> struct brw_vue_map vue_map;
> - bool has_flat_shading;
> };
>
>
> diff --git a/src/mesa/drivers/dri/i965/brw_sf_emit.c
> b/src/mesa/drivers/dri/i965/brw_sf_emit.c
> index fe05d54..5f31fa5 100644
> --- a/src/mesa/drivers/dri/i965/brw_sf_emit.c
> +++ b/src/mesa/drivers/dri/i965/brw_sf_emit.c
> @@ -162,7 +162,7 @@ static void copy_flatshaded_attributes(struct
> brw_sf_compile *c,
> int i;
>
> for (i = 0; i < c->vue_map.num_slots; i++) {
> - if (c->key.interpolation_mode.mode[i] == INTERP_MODE_FLAT) {
> + if (c->key.interp_mode[i] == INTERP_MODE_FLAT) {
> brw_MOV(p,
> get_vue_slot(c, dst, i),
> get_vue_slot(c, src, i));
> @@ -176,7 +176,7 @@ static int count_flatshaded_attributes(struct
> brw_sf_compile *c)
> int count = 0;
>
> for (i = 0; i < c->vue_map.num_slots; i++)
> - if (c->key.interpolation_mode.mode[i] == INTERP_MODE_FLAT)
> + if (c->key.interp_mode[i] == INTERP_MODE_FLAT)
> count++;
>
> return count;
> @@ -342,7 +342,7 @@ calculate_masks(struct brw_sf_compile *c,
> *pc_linear = 0;
> *pc = 0xf;
>
> - interp = c->key.interpolation_mode.mode[vert_reg_to_vue_slot(c, reg,
> 0)];
> + interp = c->key.interp_mode[vert_reg_to_vue_slot(c, reg, 0)];
> if (interp == INTERP_MODE_SMOOTH) {
> *pc_linear = 0xf;
> *pc_persp = 0xf;
> @@ -354,7 +354,7 @@ calculate_masks(struct brw_sf_compile *c,
> if (vert_reg_to_varying(c, reg, 1) != BRW_VARYING_SLOT_COUNT) {
> *pc |= 0xf0;
>
> - interp = c->key.interpolation_mode.mode[vert_reg_to_vue_slot(c,
> reg, 1)];
> + interp = c->key.interp_mode[vert_reg_to_vue_slot(c, reg, 1)];
> if (interp == INTERP_MODE_SMOOTH) {
> *pc_linear |= 0xf0;
> *pc_persp |= 0xf0;
> @@ -428,7 +428,7 @@ void brw_emit_tri_setup(struct brw_sf_compile *c, bool
> allocate)
> if (c->key.do_twoside_color)
> do_twoside_color(c);
>
> - if (c->has_flat_shading)
> + if (c->key.contains_flat_varying)
> do_flatshade_triangle(c);
>
>
> @@ -514,7 +514,7 @@ void brw_emit_line_setup(struct brw_sf_compile *c,
> bool allocate)
> invert_det(c);
> copy_z_inv_w(c);
>
> - if (c->has_flat_shading)
> + if (c->key.contains_flat_varying)
> do_flatshade_line(c);
>
> for (i = 0; i < c->nr_setup_regs; i++)
> diff --git a/src/mesa/drivers/dri/i965/brw_state.h
> b/src/mesa/drivers/dri/i965/brw_state.h
> index b42b9af..1420aab 100644
> --- a/src/mesa/drivers/dri/i965/brw_state.h
> +++ b/src/mesa/drivers/dri/i965/brw_state.h
> @@ -396,9 +396,6 @@ void gen7_enable_hw_binding_tables(struct brw_context
> *brw);
> void gen7_disable_hw_binding_tables(struct brw_context *brw);
> void gen7_reset_hw_bt_pool_offsets(struct brw_context *brw);
>
> -/* brw_interpolation_map.c */
> -void brw_setup_vue_interpolation(struct brw_context *brw);
> -
> /* brw_clip.c */
> void brw_upload_clip_prog(struct brw_context *brw);
>
> diff --git a/src/mesa/drivers/dri/i965/brw_state_upload.c
> b/src/mesa/drivers/dri/i965/brw_state_upload.c
> index 17d1b2d..17befe3 100644
> --- a/src/mesa/drivers/dri/i965/brw_state_upload.c
> +++ b/src/mesa/drivers/dri/i965/brw_state_upload.c
> @@ -734,13 +734,12 @@ brw_upload_programs(struct brw_context *brw,
> ctx->Const.MaxViewports : 1;
> }
>
> + brw_upload_wm_prog(brw);
> +
> if (brw->gen < 6) {
> - brw_setup_vue_interpolation(brw);
> brw_upload_clip_prog(brw);
> brw_upload_sf_prog(brw);
> }
> -
> - brw_upload_wm_prog(brw);
> } else if (pipeline == BRW_COMPUTE_PIPELINE) {
> brw_upload_cs_prog(brw);
> }
> diff --git a/src/mesa/drivers/dri/i965/brw_wm.c
> b/src/mesa/drivers/dri/i965/brw_wm.c
> index e65f77a..53b7d9b 100644
> --- a/src/mesa/drivers/dri/i965/brw_wm.c
> +++ b/src/mesa/drivers/dri/i965/brw_wm.c
> @@ -77,7 +77,8 @@ bool
> brw_codegen_wm_prog(struct brw_context *brw,
> struct gl_shader_program *prog,
> struct brw_fragment_program *fp,
> - struct brw_wm_prog_key *key)
> + struct brw_wm_prog_key *key,
> + struct brw_vue_map *vue_map)
> {
> const struct gen_device_info *devinfo = &brw->screen->devinfo;
> struct gl_context *ctx = &brw->ctx;
> @@ -146,8 +147,9 @@ brw_codegen_wm_prog(struct brw_context *brw,
> program = brw_compile_fs(brw->screen->compiler, brw, mem_ctx,
> key, &prog_data, fp->program.Base.nir,
> &fp->program.Base, st_index8, st_index16,
> - true, brw->use_rep_send,
> + true, brw->use_rep_send, vue_map,
> &program_size, &error_str);
> +
> if (program == NULL) {
> if (prog) {
> prog->LinkStatus = false;
> @@ -587,7 +589,8 @@ brw_upload_wm_prog(struct brw_context *brw)
> &key, sizeof(key),
> &brw->wm.base.prog_offset,
> &brw->wm.base.prog_data)) {
> - bool success = brw_codegen_wm_prog(brw, current, fp, &key);
> + bool success = brw_codegen_wm_prog(brw, current, fp, &key,
> + &brw->vue_map_geom_out);
> (void) success;
> assert(success);
> }
> @@ -641,7 +644,14 @@ brw_fs_precompile(struct gl_context *ctx,
> uint32_t old_prog_offset = brw->wm.base.prog_offset;
> struct brw_stage_prog_data *old_prog_data = brw->wm.base.prog_data;
>
> - bool success = brw_codegen_wm_prog(brw, shader_prog, bfp, &key);
> + struct brw_vue_map vue_map;
> + if (brw->gen < 6) {
> + brw_compute_vue_map(&brw->screen->devinfo, &vue_map,
> + fp->Base.nir->info->inputs_read |
> VARYING_BIT_POS,
> + false);
> + }
> +
> + bool success = brw_codegen_wm_prog(brw, shader_prog, bfp, &key,
> &vue_map);
>
> brw->wm.base.prog_offset = old_prog_offset;
> brw->wm.base.prog_data = old_prog_data;
> diff --git a/src/mesa/drivers/dri/i965/brw_wm.h
> b/src/mesa/drivers/dri/i965/brw_wm.h
> index 8fa24b1..f4f513b 100644
> --- a/src/mesa/drivers/dri/i965/brw_wm.h
> +++ b/src/mesa/drivers/dri/i965/brw_wm.h
> @@ -63,7 +63,8 @@ bool brw_color_buffer_write_enabled(struct brw_context
> *brw);
> bool brw_codegen_wm_prog(struct brw_context *brw,
> struct gl_shader_program *prog,
> struct brw_fragment_program *fp,
> - struct brw_wm_prog_key *key);
> + struct brw_wm_prog_key *key,
> + struct brw_vue_map *vue_map);
> void brw_wm_debug_recompile(struct brw_context *brw,
> struct gl_shader_program *prog,
> const struct brw_wm_prog_key *key);
> --
> 2.7.4
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
