diff options
| author | Jason Ekstrand <[email protected]> | 2015-08-14 17:25:04 -0700 |
|---|---|---|
| committer | Jason Ekstrand <[email protected]> | 2015-08-17 11:25:03 -0700 |
| commit | 6a7ca4ef2cd3f39d3b5e77051cb3f3175e9e60df (patch) | |
| tree | d5413781ac9e9ecfc22cf403fa7465d6a7cadb34 /src/gallium/auxiliary | |
| parent | b4c02253c4e1a7bc5a7a6369045210932f5de605 (diff) | |
| parent | d3e23f1ff915c01541f8df375b50b93b3da565a8 (diff) | |
Merge remote-tracking branch 'mesa-public/master' into vulkan
Diffstat (limited to 'src/gallium/auxiliary')
161 files changed, 2157 insertions, 2787 deletions
diff --git a/src/gallium/auxiliary/Makefile.am b/src/gallium/auxiliary/Makefile.am index 89c7a13e913..04f77d002c8 100644 --- a/src/gallium/auxiliary/Makefile.am +++ b/src/gallium/auxiliary/Makefile.am @@ -1,5 +1,3 @@ -AUTOMAKE_OPTIONS = subdir-objects - if HAVE_LOADER_GALLIUM SUBDIRS := pipe-loader endif @@ -10,6 +8,7 @@ include $(top_srcdir)/src/gallium/Automake.inc noinst_LTLIBRARIES = libgallium.la AM_CFLAGS = \ + -I$(top_srcdir)/src/loader \ -I$(top_builddir)/src/glsl/nir \ -I$(top_srcdir)/src/gallium/auxiliary/util \ $(GALLIUM_CFLAGS) \ diff --git a/src/gallium/auxiliary/Makefile.sources b/src/gallium/auxiliary/Makefile.sources index 62e6b94cab8..3616d885b47 100644 --- a/src/gallium/auxiliary/Makefile.sources +++ b/src/gallium/auxiliary/Makefile.sources @@ -274,7 +274,6 @@ C_SOURCES := \ util/u_simple_shaders.h \ util/u_slab.c \ util/u_slab.h \ - util/u_snprintf.c \ util/u_split_prim.h \ util/u_sse.h \ util/u_staging.c \ diff --git a/src/gallium/auxiliary/cso_cache/cso_cache.c b/src/gallium/auxiliary/cso_cache/cso_cache.c index dd56e4a154e..d36f1fbd717 100644 --- a/src/gallium/auxiliary/cso_cache/cso_cache.c +++ b/src/gallium/auxiliary/cso_cache/cso_cache.c @@ -80,7 +80,7 @@ unsigned cso_construct_key(void *item, int item_size) return hash_key((item), item_size); } -static INLINE struct cso_hash *_cso_hash_for_type(struct cso_cache *sc, enum cso_cache_type type) +static inline struct cso_hash *_cso_hash_for_type(struct cso_cache *sc, enum cso_cache_type type) { struct cso_hash *hash; hash = sc->hashes[type]; @@ -127,7 +127,7 @@ static void delete_velements(void *state, void *data) FREE(state); } -static INLINE void delete_cso(void *state, enum cso_cache_type type) +static inline void delete_cso(void *state, enum cso_cache_type type) { switch (type) { case CSO_BLEND: @@ -152,7 +152,7 @@ static INLINE void delete_cso(void *state, enum cso_cache_type type) } -static INLINE void sanitize_hash(struct cso_cache *sc, +static inline void sanitize_hash(struct cso_cache *sc, struct cso_hash *hash, enum cso_cache_type type, int max_size) @@ -162,7 +162,7 @@ static INLINE void sanitize_hash(struct cso_cache *sc, } -static INLINE void sanitize_cb(struct cso_hash *hash, enum cso_cache_type type, +static inline void sanitize_cb(struct cso_hash *hash, enum cso_cache_type type, int max_size, void *user_data) { /* if we're approach the maximum size, remove fourth of the entries diff --git a/src/gallium/auxiliary/cso_cache/cso_context.c b/src/gallium/auxiliary/cso_cache/cso_context.c index 744b00cbd92..00686d2af41 100644 --- a/src/gallium/auxiliary/cso_cache/cso_context.c +++ b/src/gallium/auxiliary/cso_cache/cso_context.c @@ -56,22 +56,8 @@ */ struct sampler_info { - struct { - void *samplers[PIPE_MAX_SAMPLERS]; - unsigned nr_samplers; - } hw; - void *samplers[PIPE_MAX_SAMPLERS]; unsigned nr_samplers; - - void *samplers_saved[PIPE_MAX_SAMPLERS]; - unsigned nr_samplers_saved; - - struct pipe_sampler_view *views[PIPE_MAX_SHADER_SAMPLER_VIEWS]; - unsigned nr_views; - - struct pipe_sampler_view *views_saved[PIPE_MAX_SHADER_SAMPLER_VIEWS]; - unsigned nr_views_saved; }; @@ -85,6 +71,15 @@ struct cso_context { boolean has_tessellation; boolean has_streamout; + struct pipe_sampler_view *fragment_views[PIPE_MAX_SHADER_SAMPLER_VIEWS]; + unsigned nr_fragment_views; + + struct pipe_sampler_view *fragment_views_saved[PIPE_MAX_SHADER_SAMPLER_VIEWS]; + unsigned nr_fragment_views_saved; + + void *fragment_samplers_saved[PIPE_MAX_SAMPLERS]; + unsigned nr_fragment_samplers_saved; + struct sampler_info samplers[PIPE_SHADER_TYPES]; struct pipe_vertex_buffer aux_vertex_buffer_current; @@ -116,9 +111,6 @@ struct cso_context { uint render_condition_mode, render_condition_mode_saved; boolean render_condition_cond, render_condition_cond_saved; - struct pipe_clip_state clip; - struct pipe_clip_state clip_saved; - struct pipe_framebuffer_state fb, fb_saved; struct pipe_viewport_state vp, vp_saved; struct pipe_blend_color blend_color; @@ -192,7 +184,7 @@ static boolean delete_vertex_elements(struct cso_context *ctx, } -static INLINE boolean delete_cso(struct cso_context *ctx, +static inline boolean delete_cso(struct cso_context *ctx, void *state, enum cso_cache_type type) { switch (type) { @@ -213,7 +205,7 @@ static INLINE boolean delete_cso(struct cso_context *ctx, return FALSE; } -static INLINE void +static inline void sanitize_hash(struct cso_hash *hash, enum cso_cache_type type, int max_size, void *user_data) { @@ -297,7 +289,7 @@ out: */ void cso_destroy_context( struct cso_context *ctx ) { - unsigned i, shader; + unsigned i; if (ctx->pipe) { ctx->pipe->set_index_buffer(ctx->pipe, NULL); @@ -347,13 +339,9 @@ void cso_destroy_context( struct cso_context *ctx ) ctx->pipe->set_stream_output_targets(ctx->pipe, 0, NULL, NULL); } - /* free sampler views for each shader stage */ - for (shader = 0; shader < Elements(ctx->samplers); shader++) { - struct sampler_info *info = &ctx->samplers[shader]; - for (i = 0; i < PIPE_MAX_SHADER_SAMPLER_VIEWS; i++) { - pipe_sampler_view_reference(&info->views[i], NULL); - pipe_sampler_view_reference(&info->views_saved[i], NULL); - } + for (i = 0; i < PIPE_MAX_SHADER_SAMPLER_VIEWS; i++) { + pipe_sampler_view_reference(&ctx->fragment_views[i], NULL); + pipe_sampler_view_reference(&ctx->fragment_views_saved[i], NULL); } util_unreference_framebuffer_state(&ctx->fb); @@ -919,47 +907,6 @@ void cso_restore_tesseval_shader(struct cso_context *ctx) ctx->tesseval_shader_saved = NULL; } -/* clip state */ - -static INLINE void -clip_state_cpy(struct pipe_clip_state *dst, - const struct pipe_clip_state *src) -{ - memcpy(dst->ucp, src->ucp, sizeof(dst->ucp)); -} - -static INLINE int -clip_state_cmp(const struct pipe_clip_state *a, - const struct pipe_clip_state *b) -{ - return memcmp(a->ucp, b->ucp, sizeof(a->ucp)); -} - -void -cso_set_clip(struct cso_context *ctx, - const struct pipe_clip_state *clip) -{ - if (clip_state_cmp(&ctx->clip, clip)) { - clip_state_cpy(&ctx->clip, clip); - ctx->pipe->set_clip_state(ctx->pipe, clip); - } -} - -void -cso_save_clip(struct cso_context *ctx) -{ - clip_state_cpy(&ctx->clip_saved, &ctx->clip); -} - -void -cso_restore_clip(struct cso_context *ctx) -{ - if (clip_state_cmp(&ctx->clip, &ctx->clip_saved)) { - clip_state_cpy(&ctx->clip, &ctx->clip_saved); - ctx->pipe->set_clip_state(ctx->pipe, &ctx->clip_saved); - } -} - enum pipe_error cso_set_vertex_elements(struct cso_context *ctx, unsigned count, @@ -1122,11 +1069,9 @@ unsigned cso_get_aux_vertex_buffer_slot(struct cso_context *ctx) /**************** fragment/vertex sampler view state *************************/ -static enum pipe_error -single_sampler(struct cso_context *ctx, - struct sampler_info *info, - unsigned idx, - const struct pipe_sampler_state *templ) +enum pipe_error +cso_single_sampler(struct cso_context *ctx, unsigned shader_stage, + unsigned idx, const struct pipe_sampler_state *templ) { void *handle = NULL; @@ -1162,24 +1107,13 @@ single_sampler(struct cso_context *ctx, } } - info->samplers[idx] = handle; - + ctx->samplers[shader_stage].samplers[idx] = handle; return PIPE_OK; } -enum pipe_error -cso_single_sampler(struct cso_context *ctx, - unsigned shader_stage, - unsigned idx, - const struct pipe_sampler_state *templ) -{ - return single_sampler(ctx, &ctx->samplers[shader_stage], idx, templ); -} - - -static void -single_sampler_done(struct cso_context *ctx, unsigned shader_stage) +void +cso_single_sampler_done(struct cso_context *ctx, unsigned shader_stage) { struct sampler_info *info = &ctx->samplers[shader_stage]; unsigned i; @@ -1191,33 +1125,8 @@ single_sampler_done(struct cso_context *ctx, unsigned shader_stage) } info->nr_samplers = i; - - if (info->hw.nr_samplers != info->nr_samplers || - memcmp(info->hw.samplers, - info->samplers, - info->nr_samplers * sizeof(void *)) != 0) - { - memcpy(info->hw.samplers, - info->samplers, - info->nr_samplers * sizeof(void *)); - - /* set remaining slots/pointers to null */ - for (i = info->nr_samplers; i < info->hw.nr_samplers; i++) - info->samplers[i] = NULL; - - ctx->pipe->bind_sampler_states(ctx->pipe, shader_stage, 0, - MAX2(info->nr_samplers, - info->hw.nr_samplers), - info->samplers); - - info->hw.nr_samplers = info->nr_samplers; - } -} - -void -cso_single_sampler_done(struct cso_context *ctx, unsigned shader_stage) -{ - single_sampler_done(ctx, shader_stage); + ctx->pipe->bind_sampler_states(ctx->pipe, shader_stage, 0, i, + info->samplers); } @@ -1240,38 +1149,42 @@ cso_set_samplers(struct cso_context *ctx, */ for (i = 0; i < nr; i++) { - temp = single_sampler(ctx, info, i, templates[i]); + temp = cso_single_sampler(ctx, shader_stage, i, templates[i]); if (temp != PIPE_OK) error = temp; } for ( ; i < info->nr_samplers; i++) { - temp = single_sampler(ctx, info, i, NULL); + temp = cso_single_sampler(ctx, shader_stage, i, NULL); if (temp != PIPE_OK) error = temp; } - single_sampler_done(ctx, shader_stage); + cso_single_sampler_done(ctx, shader_stage); return error; } void -cso_save_samplers(struct cso_context *ctx, unsigned shader_stage) +cso_save_fragment_samplers(struct cso_context *ctx) { - struct sampler_info *info = &ctx->samplers[shader_stage]; - info->nr_samplers_saved = info->nr_samplers; - memcpy(info->samplers_saved, info->samplers, sizeof(info->samplers)); + struct sampler_info *info = &ctx->samplers[PIPE_SHADER_FRAGMENT]; + + ctx->nr_fragment_samplers_saved = info->nr_samplers; + memcpy(ctx->fragment_samplers_saved, info->samplers, + sizeof(info->samplers)); } void -cso_restore_samplers(struct cso_context *ctx, unsigned shader_stage) +cso_restore_fragment_samplers(struct cso_context *ctx) { - struct sampler_info *info = &ctx->samplers[shader_stage]; - info->nr_samplers = info->nr_samplers_saved; - memcpy(info->samplers, info->samplers_saved, sizeof(info->samplers)); - single_sampler_done(ctx, shader_stage); + struct sampler_info *info = &ctx->samplers[PIPE_SHADER_FRAGMENT]; + + info->nr_samplers = ctx->nr_fragment_samplers_saved; + memcpy(info->samplers, ctx->fragment_samplers_saved, + sizeof(info->samplers)); + cso_single_sampler_done(ctx, PIPE_SHADER_FRAGMENT); } @@ -1281,71 +1194,74 @@ cso_set_sampler_views(struct cso_context *ctx, unsigned count, struct pipe_sampler_view **views) { - struct sampler_info *info = &ctx->samplers[shader_stage]; - unsigned i; - boolean any_change = FALSE; + if (shader_stage == PIPE_SHADER_FRAGMENT) { + unsigned i; + boolean any_change = FALSE; - /* reference new views */ - for (i = 0; i < count; i++) { - any_change |= info->views[i] != views[i]; - pipe_sampler_view_reference(&info->views[i], views[i]); - } - /* unref extra old views, if any */ - for (; i < info->nr_views; i++) { - any_change |= info->views[i] != NULL; - pipe_sampler_view_reference(&info->views[i], NULL); - } + /* reference new views */ + for (i = 0; i < count; i++) { + any_change |= ctx->fragment_views[i] != views[i]; + pipe_sampler_view_reference(&ctx->fragment_views[i], views[i]); + } + /* unref extra old views, if any */ + for (; i < ctx->nr_fragment_views; i++) { + any_change |= ctx->fragment_views[i] != NULL; + pipe_sampler_view_reference(&ctx->fragment_views[i], NULL); + } - /* bind the new sampler views */ - if (any_change) { - ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0, - MAX2(info->nr_views, count), - info->views); - } + /* bind the new sampler views */ + if (any_change) { + ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0, + MAX2(ctx->nr_fragment_views, count), + ctx->fragment_views); + } - info->nr_views = count; + ctx->nr_fragment_views = count; + } + else + ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0, count, views); } void -cso_save_sampler_views(struct cso_context *ctx, unsigned shader_stage) +cso_save_fragment_sampler_views(struct cso_context *ctx) { - struct sampler_info *info = &ctx->samplers[shader_stage]; unsigned i; - info->nr_views_saved = info->nr_views; + ctx->nr_fragment_views_saved = ctx->nr_fragment_views; - for (i = 0; i < info->nr_views; i++) { - assert(!info->views_saved[i]); - pipe_sampler_view_reference(&info->views_saved[i], info->views[i]); + for (i = 0; i < ctx->nr_fragment_views; i++) { + assert(!ctx->fragment_views_saved[i]); + pipe_sampler_view_reference(&ctx->fragment_views_saved[i], + ctx->fragment_views[i]); } } void -cso_restore_sampler_views(struct cso_context *ctx, unsigned shader_stage) +cso_restore_fragment_sampler_views(struct cso_context *ctx) { - struct sampler_info *info = &ctx->samplers[shader_stage]; - unsigned i, nr_saved = info->nr_views_saved; + unsigned i, nr_saved = ctx->nr_fragment_views_saved; unsigned num; for (i = 0; i < nr_saved; i++) { - pipe_sampler_view_reference(&info->views[i], NULL); + pipe_sampler_view_reference(&ctx->fragment_views[i], NULL); /* move the reference from one pointer to another */ - info->views[i] = info->views_saved[i]; - info->views_saved[i] = NULL; + ctx->fragment_views[i] = ctx->fragment_views_saved[i]; + ctx->fragment_views_saved[i] = NULL; } - for (; i < info->nr_views; i++) { - pipe_sampler_view_reference(&info->views[i], NULL); + for (; i < ctx->nr_fragment_views; i++) { + pipe_sampler_view_reference(&ctx->fragment_views[i], NULL); } - num = MAX2(info->nr_views, nr_saved); + num = MAX2(ctx->nr_fragment_views, nr_saved); /* bind the old/saved sampler views */ - ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0, num, info->views); + ctx->pipe->set_sampler_views(ctx->pipe, PIPE_SHADER_FRAGMENT, 0, num, + ctx->fragment_views); - info->nr_views = nr_saved; - info->nr_views_saved = 0; + ctx->nr_fragment_views = nr_saved; + ctx->nr_fragment_views_saved = 0; } diff --git a/src/gallium/auxiliary/cso_cache/cso_context.h b/src/gallium/auxiliary/cso_cache/cso_context.h index cc50b60c6cd..f0a27390d17 100644 --- a/src/gallium/auxiliary/cso_cache/cso_context.h +++ b/src/gallium/auxiliary/cso_cache/cso_context.h @@ -72,19 +72,17 @@ cso_set_samplers(struct cso_context *cso, const struct pipe_sampler_state **states); void -cso_save_samplers(struct cso_context *cso, unsigned shader_stage); +cso_save_fragment_samplers(struct cso_context *cso); void -cso_restore_samplers(struct cso_context *cso, unsigned shader_stage); +cso_restore_fragment_samplers(struct cso_context *cso); /* Alternate interface to support state trackers that like to modify * samplers one at a time: */ enum pipe_error -cso_single_sampler(struct cso_context *cso, - unsigned shader_stage, - unsigned count, - const struct pipe_sampler_state *states); +cso_single_sampler(struct cso_context *cso, unsigned shader_stage, + unsigned idx, const struct pipe_sampler_state *states); void cso_single_sampler_done(struct cso_context *cso, unsigned shader_stage); @@ -188,19 +186,6 @@ void cso_save_render_condition(struct cso_context *cso); void cso_restore_render_condition(struct cso_context *cso); -/* clip state */ - -void -cso_set_clip(struct cso_context *cso, - const struct pipe_clip_state *clip); - -void -cso_save_clip(struct cso_context *cso); - -void -cso_restore_clip(struct cso_context *cso); - - /* sampler view state */ void @@ -210,10 +195,10 @@ cso_set_sampler_views(struct cso_context *cso, struct pipe_sampler_view **views); void -cso_save_sampler_views(struct cso_context *cso, unsigned shader_stage); +cso_save_fragment_sampler_views(struct cso_context *ctx); void -cso_restore_sampler_views(struct cso_context *cso, unsigned shader_stage); +cso_restore_fragment_sampler_views(struct cso_context *ctx); /* constant buffers */ diff --git a/src/gallium/auxiliary/draw/draw_gs.c b/src/gallium/auxiliary/draw/draw_gs.c index a1564f93292..c827a68ea0a 100644 --- a/src/gallium/auxiliary/draw/draw_gs.c +++ b/src/gallium/auxiliary/draw/draw_gs.c @@ -45,7 +45,7 @@ /* fixme: move it from here */ #define MAX_PRIMITIVES 64 -static INLINE int +static inline int draw_gs_get_input_index(int semantic, int index, const struct tgsi_shader_info *input_info) { @@ -66,7 +66,7 @@ draw_gs_get_input_index(int semantic, int index, * the number of elements in the SOA vector. This ensures that the * throughput is optimized for the given vector instruction set. */ -static INLINE boolean +static inline boolean draw_gs_should_flush(struct draw_geometry_shader *shader) { return (shader->fetched_prim_count == shader->vector_length); diff --git a/src/gallium/auxiliary/draw/draw_llvm.c b/src/gallium/auxiliary/draw/draw_llvm.c index 90a31bc6ac0..b1e1bcbee04 100644 --- a/src/gallium/auxiliary/draw/draw_llvm.c +++ b/src/gallium/auxiliary/draw/draw_llvm.c @@ -72,7 +72,7 @@ struct draw_gs_llvm_iface { LLVMValueRef input; }; -static INLINE const struct draw_gs_llvm_iface * +static inline const struct draw_gs_llvm_iface * draw_gs_llvm_iface(const struct lp_build_tgsi_gs_iface *iface) { return (const struct draw_gs_llvm_iface *)iface; diff --git a/src/gallium/auxiliary/draw/draw_llvm.h b/src/gallium/auxiliary/draw/draw_llvm.h index d48ed721593..d153c166ead 100644 --- a/src/gallium/auxiliary/draw/draw_llvm.h +++ b/src/gallium/auxiliary/draw/draw_llvm.h @@ -350,7 +350,7 @@ struct draw_gs_llvm_variant_key PIPE_MAX_SHADER_SAMPLER_VIEWS * sizeof(struct draw_sampler_static_state)) -static INLINE size_t +static inline size_t draw_llvm_variant_key_size(unsigned nr_vertex_elements, unsigned nr_samplers) { @@ -360,7 +360,7 @@ draw_llvm_variant_key_size(unsigned nr_vertex_elements, } -static INLINE size_t +static inline size_t draw_gs_llvm_variant_key_size(unsigned nr_samplers) { return (sizeof(struct draw_gs_llvm_variant_key) + @@ -368,7 +368,7 @@ draw_gs_llvm_variant_key_size(unsigned nr_samplers) } -static INLINE struct draw_sampler_static_state * +static inline struct draw_sampler_static_state * draw_llvm_variant_key_samplers(struct draw_llvm_variant_key *key) { return (struct draw_sampler_static_state *) @@ -476,13 +476,13 @@ struct draw_llvm { }; -static INLINE struct llvm_vertex_shader * +static inline struct llvm_vertex_shader * llvm_vertex_shader(struct draw_vertex_shader *vs) { return (struct llvm_vertex_shader *)vs; } -static INLINE struct llvm_geometry_shader * +static inline struct llvm_geometry_shader * llvm_geometry_shader(struct draw_geometry_shader *gs) { return (struct llvm_geometry_shader *)gs; diff --git a/src/gallium/auxiliary/draw/draw_pipe.h b/src/gallium/auxiliary/draw/draw_pipe.h index 35273330d13..e69dcbded0e 100644 --- a/src/gallium/auxiliary/draw/draw_pipe.h +++ b/src/gallium/auxiliary/draw/draw_pipe.h @@ -115,7 +115,7 @@ void draw_unfilled_prepare_outputs(struct draw_context *context, * \param idx index into stage's tmp[] array to put the copy (dest) * \return pointer to the copied vertex */ -static INLINE struct vertex_header * +static inline struct vertex_header * dup_vert( struct draw_stage *stage, const struct vertex_header *vert, unsigned idx ) diff --git a/src/gallium/auxiliary/draw/draw_pipe_aaline.c b/src/gallium/auxiliary/draw/draw_pipe_aaline.c index 936046ea5f5..85d24b7a6a1 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_aaline.c +++ b/src/gallium/auxiliary/draw/draw_pipe_aaline.c @@ -511,7 +511,7 @@ bind_aaline_fragment_shader(struct aaline_stage *aaline) -static INLINE struct aaline_stage * +static inline struct aaline_stage * aaline_stage( struct draw_stage *stage ) { return (struct aaline_stage *) stage; diff --git a/src/gallium/auxiliary/draw/draw_pipe_aapoint.c b/src/gallium/auxiliary/draw/draw_pipe_aapoint.c index 7feb49ae934..3918923296d 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_aapoint.c +++ b/src/gallium/auxiliary/draw/draw_pipe_aapoint.c @@ -427,7 +427,7 @@ bind_aapoint_fragment_shader(struct aapoint_stage *aapoint) -static INLINE struct aapoint_stage * +static inline struct aapoint_stage * aapoint_stage( struct draw_stage *stage ) { return (struct aapoint_stage *) stage; diff --git a/src/gallium/auxiliary/draw/draw_pipe_clip.c b/src/gallium/auxiliary/draw/draw_pipe_clip.c index e1e7dcc6f63..c22758bc702 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_clip.c +++ b/src/gallium/auxiliary/draw/draw_pipe_clip.c @@ -70,12 +70,12 @@ struct clip_stage { /** Cast wrapper */ -static INLINE struct clip_stage *clip_stage( struct draw_stage *stage ) +static inline struct clip_stage *clip_stage( struct draw_stage *stage ) { return (struct clip_stage *)stage; } -static INLINE unsigned +static inline unsigned draw_viewport_index(struct draw_context *draw, const struct vertex_header *leading_vertex) { @@ -210,7 +210,7 @@ static void interp( const struct clip_stage *clip, * true, otherwise returns false. * Triangle is considered null/empty if it's area is qual to zero. */ -static INLINE boolean +static inline boolean is_tri_null(struct draw_context *draw, const struct prim_header *header) { const unsigned pos_attr = draw_current_shader_position_output(draw); @@ -322,7 +322,7 @@ static void emit_poly( struct draw_stage *stage, } -static INLINE float +static inline float dot4(const float *a, const float *b) { return (a[0] * b[0] + @@ -336,7 +336,7 @@ dot4(const float *a, const float *b) * it first checks if the shader provided a clip distance, otherwise * it works out the value using the clipvertex */ -static INLINE float getclipdist(const struct clip_stage *clipper, +static inline float getclipdist(const struct clip_stage *clipper, struct vertex_header *vert, int plane_idx) { diff --git a/src/gallium/auxiliary/draw/draw_pipe_cull.c b/src/gallium/auxiliary/draw/draw_pipe_cull.c index fa344089a8a..fc8293bd128 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_cull.c +++ b/src/gallium/auxiliary/draw/draw_pipe_cull.c @@ -46,12 +46,12 @@ struct cull_stage { }; -static INLINE struct cull_stage *cull_stage( struct draw_stage *stage ) +static inline struct cull_stage *cull_stage( struct draw_stage *stage ) { return (struct cull_stage *)stage; } -static INLINE boolean +static inline boolean cull_distance_is_out(float dist) { return (dist < 0.0f) || util_is_inf_or_nan(dist); diff --git a/src/gallium/auxiliary/draw/draw_pipe_flatshade.c b/src/gallium/auxiliary/draw/draw_pipe_flatshade.c index 59e33b472f4..0ea740861d6 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_flatshade.c +++ b/src/gallium/auxiliary/draw/draw_pipe_flatshade.c @@ -47,7 +47,7 @@ struct flat_stage }; -static INLINE struct flat_stage * +static inline struct flat_stage * flat_stage(struct draw_stage *stage) { return (struct flat_stage *) stage; @@ -55,7 +55,7 @@ flat_stage(struct draw_stage *stage) /** Copy all the constant attributes from 'src' vertex to 'dst' vertex */ -static INLINE void copy_flats( struct draw_stage *stage, +static inline void copy_flats( struct draw_stage *stage, struct vertex_header *dst, const struct vertex_header *src ) { @@ -70,7 +70,7 @@ static INLINE void copy_flats( struct draw_stage *stage, /** Copy all the color attributes from src vertex to dst0 & dst1 vertices */ -static INLINE void copy_flats2( struct draw_stage *stage, +static inline void copy_flats2( struct draw_stage *stage, struct vertex_header *dst0, struct vertex_header *dst1, const struct vertex_header *src ) diff --git a/src/gallium/auxiliary/draw/draw_pipe_offset.c b/src/gallium/auxiliary/draw/draw_pipe_offset.c index b25dd21fd4d..5e0d8ce793d 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_offset.c +++ b/src/gallium/auxiliary/draw/draw_pipe_offset.c @@ -49,7 +49,7 @@ struct offset_stage { -static INLINE struct offset_stage *offset_stage( struct draw_stage *stage ) +static inline struct offset_stage *offset_stage( struct draw_stage *stage ) { return (struct offset_stage *) stage; } diff --git a/src/gallium/auxiliary/draw/draw_pipe_pstipple.c b/src/gallium/auxiliary/draw/draw_pipe_pstipple.c index 445f195e59c..186b4cb4935 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_pstipple.c +++ b/src/gallium/auxiliary/draw/draw_pipe_pstipple.c @@ -462,7 +462,7 @@ bind_pstip_fragment_shader(struct pstip_stage *pstip) } -static INLINE struct pstip_stage * +static inline struct pstip_stage * pstip_stage( struct draw_stage *stage ) { return (struct pstip_stage *) stage; diff --git a/src/gallium/auxiliary/draw/draw_pipe_stipple.c b/src/gallium/auxiliary/draw/draw_pipe_stipple.c index 476c011b9a0..381aa41530b 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_stipple.c +++ b/src/gallium/auxiliary/draw/draw_pipe_stipple.c @@ -53,7 +53,7 @@ struct stipple_stage { }; -static INLINE struct stipple_stage * +static inline struct stipple_stage * stipple_stage(struct draw_stage *stage) { return (struct stipple_stage *) stage; @@ -108,7 +108,7 @@ emit_segment(struct draw_stage *stage, struct prim_header *header, } -static INLINE unsigned +static inline unsigned stipple_test(int counter, ushort pattern, int factor) { int b = (counter / factor) & 0xf; diff --git a/src/gallium/auxiliary/draw/draw_pipe_twoside.c b/src/gallium/auxiliary/draw/draw_pipe_twoside.c index 8148f6b4569..7f958d9b985 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_twoside.c +++ b/src/gallium/auxiliary/draw/draw_pipe_twoside.c @@ -43,7 +43,7 @@ struct twoside_stage { }; -static INLINE struct twoside_stage *twoside_stage( struct draw_stage *stage ) +static inline struct twoside_stage *twoside_stage( struct draw_stage *stage ) { return (struct twoside_stage *)stage; } @@ -51,7 +51,7 @@ static INLINE struct twoside_stage *twoside_stage( struct draw_stage *stage ) /** * Copy back color(s) to front color(s). */ -static INLINE struct vertex_header * +static inline struct vertex_header * copy_bfc( struct twoside_stage *twoside, const struct vertex_header *v, unsigned idx ) diff --git a/src/gallium/auxiliary/draw/draw_pipe_unfilled.c b/src/gallium/auxiliary/draw/draw_pipe_unfilled.c index 51fbdb97ae8..8e6435cdbb4 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_unfilled.c +++ b/src/gallium/auxiliary/draw/draw_pipe_unfilled.c @@ -53,7 +53,7 @@ struct unfilled_stage { }; -static INLINE struct unfilled_stage *unfilled_stage( struct draw_stage *stage ) +static inline struct unfilled_stage *unfilled_stage( struct draw_stage *stage ) { return (struct unfilled_stage *)stage; } diff --git a/src/gallium/auxiliary/draw/draw_pipe_vbuf.c b/src/gallium/auxiliary/draw/draw_pipe_vbuf.c index e0e32dd9bbe..5cc866d7eee 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_vbuf.c +++ b/src/gallium/auxiliary/draw/draw_pipe_vbuf.c @@ -85,7 +85,7 @@ struct vbuf_stage { /** * Basically a cast wrapper. */ -static INLINE struct vbuf_stage * +static inline struct vbuf_stage * vbuf_stage( struct draw_stage *stage ) { assert(stage); @@ -97,7 +97,7 @@ static void vbuf_flush_vertices( struct vbuf_stage *vbuf ); static void vbuf_alloc_vertices( struct vbuf_stage *vbuf ); -static INLINE boolean +static inline boolean overflow( void *map, void *ptr, unsigned bytes, unsigned bufsz ) { unsigned long used = (unsigned long) ((char *)ptr - (char *)map); @@ -105,7 +105,7 @@ overflow( void *map, void *ptr, unsigned bytes, unsigned bufsz ) } -static INLINE void +static inline void check_space( struct vbuf_stage *vbuf, unsigned nr ) { if (vbuf->nr_vertices + nr > vbuf->max_vertices || @@ -126,7 +126,7 @@ check_space( struct vbuf_stage *vbuf, unsigned nr ) * have a couple of slots at the beginning (1-dword header, 4-dword * clip pos) that we ignore here. We only use the vertex->data[] fields. */ -static INLINE ushort +static inline ushort emit_vertex( struct vbuf_stage *vbuf, struct vertex_header *vertex ) { diff --git a/src/gallium/auxiliary/draw/draw_pipe_wide_line.c b/src/gallium/auxiliary/draw/draw_pipe_wide_line.c index 6c57d5c1e3e..38ac11a9adf 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_wide_line.c +++ b/src/gallium/auxiliary/draw/draw_pipe_wide_line.c @@ -45,7 +45,7 @@ struct wideline_stage { -static INLINE struct wideline_stage *wideline_stage( struct draw_stage *stage ) +static inline struct wideline_stage *wideline_stage( struct draw_stage *stage ) { return (struct wideline_stage *)stage; } diff --git a/src/gallium/auxiliary/draw/draw_pipe_wide_point.c b/src/gallium/auxiliary/draw/draw_pipe_wide_point.c index 05beba8cd97..348b0e93bbc 100644 --- a/src/gallium/auxiliary/draw/draw_pipe_wide_point.c +++ b/src/gallium/auxiliary/draw/draw_pipe_wide_point.c @@ -83,7 +83,7 @@ struct widepoint_stage { -static INLINE struct widepoint_stage * +static inline struct widepoint_stage * widepoint_stage( struct draw_stage *stage ) { return (struct widepoint_stage *)stage; diff --git a/src/gallium/auxiliary/draw/draw_private.h b/src/gallium/auxiliary/draw/draw_private.h index 7b893cb2692..0ad94bb031f 100644 --- a/src/gallium/auxiliary/draw/draw_private.h +++ b/src/gallium/auxiliary/draw/draw_private.h @@ -494,7 +494,7 @@ void draw_update_viewport_flags(struct draw_context *draw); * Return index of the given viewport clamping it * to be between 0 <= and < PIPE_MAX_VIEWPORTS */ -static INLINE unsigned +static inline unsigned draw_clamp_viewport_idx(int idx) { return ((PIPE_MAX_VIEWPORTS > idx && idx >= 0) ? idx : 0); @@ -505,7 +505,7 @@ draw_clamp_viewport_idx(int idx) * overflows then it returns the value from * the overflow_value variable. */ -static INLINE unsigned +static inline unsigned draw_overflow_uadd(unsigned a, unsigned b, unsigned overflow_value) { diff --git a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c index 5af845ff938..ffec863ae6f 100644 --- a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c +++ b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c @@ -54,7 +54,7 @@ struct fetch_pipeline_middle_end { /** cast wrapper */ -static INLINE struct fetch_pipeline_middle_end * +static inline struct fetch_pipeline_middle_end * fetch_pipeline_middle_end(struct draw_pt_middle_end *middle) { return (struct fetch_pipeline_middle_end *) middle; diff --git a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline_llvm.c b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline_llvm.c index d17d6959b44..e42c4af0e70 100644 --- a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline_llvm.c +++ b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline_llvm.c @@ -60,7 +60,7 @@ struct llvm_middle_end { /** cast wrapper */ -static INLINE struct llvm_middle_end * +static inline struct llvm_middle_end * llvm_middle_end(struct draw_pt_middle_end *middle) { return (struct llvm_middle_end *) middle; diff --git a/src/gallium/auxiliary/draw/draw_pt_post_vs.c b/src/gallium/auxiliary/draw/draw_pt_post_vs.c index 71a7d3918e9..f0d5e0f5656 100644 --- a/src/gallium/auxiliary/draw/draw_pt_post_vs.c +++ b/src/gallium/auxiliary/draw/draw_pt_post_vs.c @@ -53,7 +53,7 @@ struct pt_post_vs { const struct draw_prim_info *prim_info ); }; -static INLINE void +static inline void initialize_vertex_header(struct vertex_header *header) { header->clipmask = 0; @@ -62,7 +62,7 @@ initialize_vertex_header(struct vertex_header *header) header->vertex_id = UNDEFINED_VERTEX_ID; } -static INLINE float +static inline float dot4(const float *a, const float *b) { return (a[0]*b[0] + diff --git a/src/gallium/auxiliary/draw/draw_pt_so_emit.c b/src/gallium/auxiliary/draw/draw_pt_so_emit.c index 91e67c0840d..20de26fd08a 100644 --- a/src/gallium/auxiliary/draw/draw_pt_so_emit.c +++ b/src/gallium/auxiliary/draw/draw_pt_so_emit.c @@ -65,7 +65,7 @@ draw_so_info(const struct draw_context *draw) return state; } -static INLINE boolean +static inline boolean draw_has_so(const struct draw_context *draw) { const struct pipe_stream_output_info *state = draw_so_info(draw); diff --git a/src/gallium/auxiliary/draw/draw_pt_vsplit.c b/src/gallium/auxiliary/draw/draw_pt_vsplit.c index 8098adea61f..8d448f92a26 100644 --- a/src/gallium/auxiliary/draw/draw_pt_vsplit.c +++ b/src/gallium/auxiliary/draw/draw_pt_vsplit.c @@ -84,7 +84,7 @@ vsplit_flush_cache(struct vsplit_frontend *vsplit, unsigned flags) /** * Add a fetch element and add it to the draw elements. */ -static INLINE void +static inline void vsplit_add_cache(struct vsplit_frontend *vsplit, unsigned fetch, unsigned ofbias) { unsigned hash; @@ -111,7 +111,7 @@ vsplit_add_cache(struct vsplit_frontend *vsplit, unsigned fetch, unsigned ofbias * The value is checked for overflows (both integer overflows * and the elements array overflow). */ -static INLINE unsigned +static inline unsigned vsplit_get_base_idx(struct vsplit_frontend *vsplit, unsigned start, unsigned fetch, unsigned *ofbit) { @@ -137,7 +137,7 @@ vsplit_get_base_idx(struct vsplit_frontend *vsplit, * index, plus the element bias, clamped to maximum elememt * index if that addition overflows. */ -static INLINE unsigned +static inline unsigned vsplit_get_bias_idx(struct vsplit_frontend *vsplit, int idx, int bias, unsigned *ofbias) { @@ -170,7 +170,7 @@ vsplit_get_bias_idx(struct vsplit_frontend *vsplit, elt_idx = vsplit_get_base_idx(vsplit, start, fetch, &ofbit); \ elt_idx = vsplit_get_bias_idx(vsplit, ofbit ? 0 : DRAW_GET_IDX(elts, elt_idx), elt_bias, &ofbias) -static INLINE void +static inline void vsplit_add_cache_ubyte(struct vsplit_frontend *vsplit, const ubyte *elts, unsigned start, unsigned fetch, int elt_bias) { @@ -179,7 +179,7 @@ vsplit_add_cache_ubyte(struct vsplit_frontend *vsplit, const ubyte *elts, vsplit_add_cache(vsplit, elt_idx, ofbias); } -static INLINE void +static inline void vsplit_add_cache_ushort(struct vsplit_frontend *vsplit, const ushort *elts, unsigned start, unsigned fetch, int elt_bias) { @@ -193,7 +193,7 @@ vsplit_add_cache_ushort(struct vsplit_frontend *vsplit, const ushort *elts, * Add a fetch element and add it to the draw elements. The fetch element is * in full range (uint). */ -static INLINE void +static inline void vsplit_add_cache_uint(struct vsplit_frontend *vsplit, const uint *elts, unsigned start, unsigned fetch, int elt_bias) { diff --git a/src/gallium/auxiliary/draw/draw_pt_vsplit_tmp.h b/src/gallium/auxiliary/draw/draw_pt_vsplit_tmp.h index 0f7a3cdc012..0afabb01398 100644 --- a/src/gallium/auxiliary/draw/draw_pt_vsplit_tmp.h +++ b/src/gallium/auxiliary/draw/draw_pt_vsplit_tmp.h @@ -129,7 +129,7 @@ CONCAT(vsplit_primitive_, ELT_TYPE)(struct vsplit_frontend *vsplit, * When spoken is TRUE, ispoken replaces istart; When close is TRUE, iclose is * appended. */ -static INLINE void +static inline void CONCAT(vsplit_segment_cache_, ELT_TYPE)(struct vsplit_frontend *vsplit, unsigned flags, unsigned istart, unsigned icount, diff --git a/src/gallium/auxiliary/draw/draw_vertex.h b/src/gallium/auxiliary/draw/draw_vertex.h index b4178d6a6c5..ee11d2f9276 100644 --- a/src/gallium/auxiliary/draw/draw_vertex.h +++ b/src/gallium/auxiliary/draw/draw_vertex.h @@ -91,13 +91,13 @@ struct vertex_info } attrib[PIPE_MAX_SHADER_OUTPUTS]; }; -static INLINE size_t +static inline size_t draw_vinfo_size( const struct vertex_info *a ) { return offsetof(const struct vertex_info, attrib[a->num_attribs]); } -static INLINE int +static inline int draw_vinfo_compare( const struct vertex_info *a, const struct vertex_info *b ) { @@ -105,7 +105,7 @@ draw_vinfo_compare( const struct vertex_info *a, return memcmp( a, b, sizea ); } -static INLINE void +static inline void draw_vinfo_copy( struct vertex_info *dst, const struct vertex_info *src ) { @@ -121,7 +121,7 @@ draw_vinfo_copy( struct vertex_info *dst, * corresponds to this attribute. * \return slot in which the attribute was added */ -static INLINE uint +static inline uint draw_emit_vertex_attr(struct vertex_info *vinfo, enum attrib_emit emit, enum interp_mode interp, /* only used by softpipe??? */ @@ -150,7 +150,7 @@ void draw_dump_emitted_vertex(const struct vertex_info *vinfo, const uint8_t *data); -static INLINE enum pipe_format draw_translate_vinfo_format(enum attrib_emit emit) +static inline enum pipe_format draw_translate_vinfo_format(enum attrib_emit emit) { switch (emit) { case EMIT_OMIT: @@ -174,7 +174,7 @@ static INLINE enum pipe_format draw_translate_vinfo_format(enum attrib_emit emit } } -static INLINE unsigned draw_translate_vinfo_size(enum attrib_emit emit) +static inline unsigned draw_translate_vinfo_size(enum attrib_emit emit) { switch (emit) { case EMIT_OMIT: diff --git a/src/gallium/auxiliary/draw/draw_vs.h b/src/gallium/auxiliary/draw/draw_vs.h index 1d54e7ef298..24b29e70dd9 100644 --- a/src/gallium/auxiliary/draw/draw_vs.h +++ b/src/gallium/auxiliary/draw/draw_vs.h @@ -191,12 +191,12 @@ draw_vs_create_variant_generic( struct draw_vertex_shader *vs, -static INLINE int draw_vs_variant_keysize( const struct draw_vs_variant_key *key ) +static inline int draw_vs_variant_keysize( const struct draw_vs_variant_key *key ) { return 2 * sizeof(int) + key->nr_elements * sizeof(struct draw_variant_element); } -static INLINE int draw_vs_variant_key_compare( const struct draw_vs_variant_key *a, +static inline int draw_vs_variant_key_compare( const struct draw_vs_variant_key *a, const struct draw_vs_variant_key *b ) { int keysize = draw_vs_variant_keysize(a); diff --git a/src/gallium/auxiliary/gallivm/lp_bld_arit.c b/src/gallium/auxiliary/gallivm/lp_bld_arit.c index 9daa93eec3e..50ae192325b 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_arit.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_arit.c @@ -1135,7 +1135,7 @@ lp_build_div(struct lp_build_context *bld, * * @sa http://www.stereopsis.com/doubleblend.html */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_lerp_simple(struct lp_build_context *bld, LLVMValueRef x, LLVMValueRef v0, @@ -1674,7 +1674,7 @@ enum lp_build_round_mode * NOTE: In the SSE4.1's nearest mode, if two values are equally close, the * result is the even value. That is, rounding 2.5 will be 2.0, and not 3.0. */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_round_sse41(struct lp_build_context *bld, LLVMValueRef a, enum lp_build_round_mode mode) @@ -1717,7 +1717,7 @@ lp_build_round_sse41(struct lp_build_context *bld, args[2] = LLVMConstInt(i32t, mode, 0); res = lp_build_intrinsic(builder, intrinsic, - vec_type, args, Elements(args)); + vec_type, args, Elements(args), 0); res = LLVMBuildExtractElement(builder, res, index0, ""); } @@ -1761,7 +1761,7 @@ lp_build_round_sse41(struct lp_build_context *bld, } -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_iround_nearest_sse2(struct lp_build_context *bld, LLVMValueRef a) { @@ -1817,7 +1817,7 @@ lp_build_iround_nearest_sse2(struct lp_build_context *bld, /* */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_round_altivec(struct lp_build_context *bld, LLVMValueRef a, enum lp_build_round_mode mode) @@ -1851,7 +1851,7 @@ lp_build_round_altivec(struct lp_build_context *bld, return lp_build_intrinsic_unary(builder, intrinsic, bld->vec_type, a); } -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_round_arch(struct lp_build_context *bld, LLVMValueRef a, enum lp_build_round_mode mode) @@ -1997,6 +1997,12 @@ lp_build_floor(struct lp_build_context *bld, LLVMTypeRef int_vec_type = bld->int_vec_type; LLVMTypeRef vec_type = bld->vec_type; + if (type.width != 32) { + char intrinsic[32]; + util_snprintf(intrinsic, sizeof intrinsic, "llvm.floor.v%uf%u", type.length, type.width); + return lp_build_intrinsic_unary(builder, intrinsic, vec_type, a); + } + assert(type.width == 32); /* might want to handle doubles at some point */ inttype = type; @@ -2066,6 +2072,12 @@ lp_build_ceil(struct lp_build_context *bld, LLVMTypeRef int_vec_type = bld->int_vec_type; LLVMTypeRef vec_type = bld->vec_type; + if (type.width != 32) { + char intrinsic[32]; + util_snprintf(intrinsic, sizeof intrinsic, "llvm.ceil.v%uf%u", type.length, type.width); + return lp_build_intrinsic_unary(builder, intrinsic, vec_type, a); + } + assert(type.width == 32); /* might want to handle doubles at some point */ inttype = type; @@ -2427,7 +2439,7 @@ lp_build_sqrt(struct lp_build_context *bld, * - http://en.wikipedia.org/wiki/Division_(digital)#Newton.E2.80.93Raphson_division * - http://softwarecommunity.intel.com/articles/eng/1818.htm */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_rcp_refine(struct lp_build_context *bld, LLVMValueRef a, LLVMValueRef rcp_a) @@ -2512,7 +2524,7 @@ lp_build_rcp(struct lp_build_context *bld, * * See also Intel 64 and IA-32 Architectures Optimization Manual. */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_rsqrt_refine(struct lp_build_context *bld, LLVMValueRef a, LLVMValueRef rsqrt_a) @@ -3535,7 +3547,7 @@ lp_build_fpstate_get(struct gallivm_state *gallivm) lp_build_intrinsic(builder, "llvm.x86.sse.stmxcsr", LLVMVoidTypeInContext(gallivm->context), - &mxcsr_ptr8, 1); + &mxcsr_ptr8, 1, 0); return mxcsr_ptr; } return 0; @@ -3582,6 +3594,6 @@ lp_build_fpstate_set(struct gallivm_state *gallivm, lp_build_intrinsic(builder, "llvm.x86.sse.ldmxcsr", LLVMVoidTypeInContext(gallivm->context), - &mxcsr_ptr, 1); + &mxcsr_ptr, 1, 0); } } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_const.h b/src/gallium/auxiliary/gallivm/lp_bld_const.h index b17c41931f4..a4c3bf0977a 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_const.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_const.h @@ -120,14 +120,14 @@ lp_build_const_mask_aos_swizzled(struct gallivm_state *gallivm, const unsigned char *swizzle); -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_const_int32(struct gallivm_state *gallivm, int i) { return LLVMConstInt(LLVMInt32TypeInContext(gallivm->context), i, 0); } -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_const_float(struct gallivm_state *gallivm, float x) { return LLVMConstReal(LLVMFloatTypeInContext(gallivm->context), x); @@ -135,7 +135,7 @@ lp_build_const_float(struct gallivm_state *gallivm, float x) /** Return constant-valued pointer to int */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_const_int_pointer(struct gallivm_state *gallivm, const void *ptr) { LLVMTypeRef int_type; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp b/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp index 405e6486f7a..7283e2f162f 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp +++ b/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp @@ -61,6 +61,7 @@ lp_check_alignment(const void *ptr, unsigned alignment) return ((uintptr_t)ptr & (alignment - 1)) == 0; } +#if (defined(PIPE_OS_WINDOWS) && !defined(PIPE_CC_MSVC)) || defined(PIPE_OS_EMBEDDED) class raw_debug_ostream : public llvm::raw_ostream @@ -91,6 +92,7 @@ raw_debug_ostream::write_impl(const char *Ptr, size_t Size) } } +#endif extern "C" const char * lp_get_module_id(LLVMModuleRef module) @@ -123,7 +125,7 @@ lp_debug_dump_value(LLVMValueRef value) * - http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html */ static size_t -disassemble(const void* func, llvm::raw_ostream & Out) +disassemble(const void* func) { const uint8_t *bytes = (const uint8_t *)func; @@ -141,7 +143,8 @@ disassemble(const void* func, llvm::raw_ostream & Out) char outline[1024]; if (!D) { - Out << "error: couldn't create disassembler for triple " << Triple << "\n"; + _debug_printf("error: couldn't create disassembler for triple %s\n", + Triple.c_str()); return 0; } @@ -155,13 +158,13 @@ disassemble(const void* func, llvm::raw_ostream & Out) * so that between runs. */ - Out << llvm::format("%6lu:\t", (unsigned long)pc); + _debug_printf("%6lu:\t", (unsigned long)pc); Size = LLVMDisasmInstruction(D, (uint8_t *)bytes + pc, extent - pc, 0, outline, sizeof outline); if (!Size) { - Out << "invalid\n"; + _debug_printf("invalid\n"); pc += 1; break; } @@ -173,10 +176,10 @@ disassemble(const void* func, llvm::raw_ostream & Out) if (0) { unsigned i; for (i = 0; i < Size; ++i) { - Out << llvm::format("%02x ", bytes[pc + i]); + _debug_printf("%02x ", bytes[pc + i]); } for (; i < 16; ++i) { - Out << " "; + _debug_printf(" "); } } @@ -184,9 +187,9 @@ disassemble(const void* func, llvm::raw_ostream & Out) * Print the instruction. */ - Out << outline; + _debug_printf("%*s", Size, outline); - Out << "\n"; + _debug_printf("\n"); /* * Stop disassembling on return statements, if there is no record of a @@ -206,13 +209,12 @@ disassemble(const void* func, llvm::raw_ostream & Out) pc += Size; if (pc >= extent) { - Out << "disassembly larger than " << extent << "bytes, aborting\n"; + _debug_printf("disassembly larger than %ull bytes, aborting\n", extent); break; } } - Out << "\n"; - Out.flush(); + _debug_printf("\n"); LLVMDisasmDispose(D); @@ -229,9 +231,8 @@ disassemble(const void* func, llvm::raw_ostream & Out) extern "C" void lp_disassemble(LLVMValueRef func, const void *code) { - raw_debug_ostream Out; - Out << LLVMGetValueName(func) << ":\n"; - disassemble(code, Out); + _debug_printf("%s:\n", LLVMGetValueName(func)); + disassemble(code); } @@ -273,7 +274,7 @@ lp_profile(LLVMValueRef func, const void *code) unsigned long addr = (uintptr_t)code; llvm::raw_fd_ostream Out(perf_asm_fd, false); Out << symbol << ":\n"; - unsigned long size = disassemble(code, Out); + unsigned long size = disassemble(code); fprintf(perf_map_file, "%lx %lx %s\n", addr, size, symbol); fflush(perf_map_file); } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_debug.h b/src/gallium/auxiliary/gallivm/lp_bld_debug.h index 321e09d56b9..375ba6cb5ff 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_debug.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_debug.h @@ -59,7 +59,7 @@ extern unsigned gallivm_debug; #endif -static INLINE void +static inline void lp_build_name(LLVMValueRef val, const char *format, ...) { #ifdef DEBUG diff --git a/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c b/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c index efe71704c3a..ddf3ad1dfc6 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c @@ -95,7 +95,7 @@ lp_build_format_swizzle_aos(const struct util_format_description *desc, /** * Whether the format matches the vector type, apart of swizzles. */ -static INLINE boolean +static inline boolean format_matches_type(const struct util_format_description *desc, struct lp_type type) { @@ -146,7 +146,7 @@ format_matches_type(const struct util_format_description *desc, * * @return XYZW in a float[4] or ubyte[4] or ushort[4] vector. */ -static INLINE LLVMValueRef +static inline LLVMValueRef lp_build_unpack_arith_rgba_aos(struct gallivm_state *gallivm, const struct util_format_description *desc, LLVMValueRef packed) diff --git a/src/gallium/auxiliary/gallivm/lp_bld_format_yuv.c b/src/gallium/auxiliary/gallivm/lp_bld_format_yuv.c index 4f5a45c6a3d..fa0e8b656bb 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_format_yuv.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_format_yuv.c @@ -212,7 +212,7 @@ yuyv_to_yuv_soa(struct gallivm_state *gallivm, } -static INLINE void +static inline void yuv_to_rgb_soa(struct gallivm_state *gallivm, unsigned n, LLVMValueRef y, LLVMValueRef u, LLVMValueRef v, diff --git a/src/gallium/auxiliary/gallivm/lp_bld_init.c b/src/gallium/auxiliary/gallivm/lp_bld_init.c index 384ea864081..017d0752060 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_init.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_init.c @@ -106,7 +106,6 @@ enum LLVM_CodeGenOpt_Level { static boolean create_pass_manager(struct gallivm_state *gallivm) { - char *td_str; assert(!gallivm->passmgr); assert(gallivm->target); @@ -122,10 +121,29 @@ create_pass_manager(struct gallivm_state *gallivm) // Old versions of LLVM get the DataLayout from the pass manager. LLVMAddTargetData(gallivm->target, gallivm->passmgr); - // New ones from the Module. - td_str = LLVMCopyStringRepOfTargetData(gallivm->target); - LLVMSetDataLayout(gallivm->module, td_str); - free(td_str); + /* Setting the module's DataLayout to an empty string will cause the + * ExecutionEngine to copy to the DataLayout string from its target + * machine to the module. As of LLVM 3.8 the module and the execution + * engine are required to have the same DataLayout. + * + * TODO: This is just a temporary work-around. The correct solution is + * for gallivm_init_state() to create a TargetMachine and pull the + * DataLayout from there. Currently, the TargetMachine used by llvmpipe + * is being implicitly created by the EngineBuilder in + * lp_build_create_jit_compiler_for_module() + */ + +#if HAVE_LLVM < 0x0308 + { + char *td_str; + // New ones from the Module. + td_str = LLVMCopyStringRepOfTargetData(gallivm->target); + LLVMSetDataLayout(gallivm->module, td_str); + free(td_str); + } +#else + LLVMSetDataLayout(gallivm->module, ""); +#endif if ((gallivm_debug & GALLIVM_DEBUG_NO_OPT) == 0) { /* These are the passes currently listed in llvm-c/Transforms/Scalar.h, diff --git a/src/gallium/auxiliary/gallivm/lp_bld_intr.c b/src/gallium/auxiliary/gallivm/lp_bld_intr.c index 2bf1211bcd7..30f4863ec44 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_intr.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_intr.c @@ -81,7 +81,8 @@ lp_build_intrinsic(LLVMBuilderRef builder, const char *name, LLVMTypeRef ret_type, LLVMValueRef *args, - unsigned num_args) + unsigned num_args, + LLVMAttribute attr) { LLVMModuleRef module = LLVMGetGlobalParent(LLVMGetBasicBlockParent(LLVMGetInsertBlock(builder))); LLVMValueRef function; @@ -99,6 +100,9 @@ lp_build_intrinsic(LLVMBuilderRef builder, } function = lp_declare_intrinsic(module, name, ret_type, arg_types, num_args); + + if (attr) + LLVMAddFunctionAttr(function, attr); } return LLVMBuildCall(builder, function, args, num_args, ""); @@ -111,7 +115,7 @@ lp_build_intrinsic_unary(LLVMBuilderRef builder, LLVMTypeRef ret_type, LLVMValueRef a) { - return lp_build_intrinsic(builder, name, ret_type, &a, 1); + return lp_build_intrinsic(builder, name, ret_type, &a, 1, 0); } @@ -127,7 +131,7 @@ lp_build_intrinsic_binary(LLVMBuilderRef builder, args[0] = a; args[1] = b; - return lp_build_intrinsic(builder, name, ret_type, args, 2); + return lp_build_intrinsic(builder, name, ret_type, args, 2, 0); } @@ -242,7 +246,7 @@ lp_build_intrinsic_map(struct gallivm_state *gallivm, LLVMValueRef res_elem; for(j = 0; j < num_args; ++j) arg_elems[j] = LLVMBuildExtractElement(builder, args[j], index, ""); - res_elem = lp_build_intrinsic(builder, name, ret_elem_type, arg_elems, num_args); + res_elem = lp_build_intrinsic(builder, name, ret_elem_type, arg_elems, num_args, 0); res = LLVMBuildInsertElement(builder, res, res_elem, index, ""); } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_intr.h b/src/gallium/auxiliary/gallivm/lp_bld_intr.h index 38c5c29c980..a54b367961a 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_intr.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_intr.h @@ -59,7 +59,8 @@ lp_build_intrinsic(LLVMBuilderRef builder, const char *name, LLVMTypeRef ret_type, LLVMValueRef *args, - unsigned num_args); + unsigned num_args, + LLVMAttribute attr); LLVMValueRef diff --git a/src/gallium/auxiliary/gallivm/lp_bld_limits.h b/src/gallium/auxiliary/gallivm/lp_bld_limits.h index db503514881..571c615f9f8 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_limits.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_limits.h @@ -88,7 +88,7 @@ * actually try to allocate the maximum and run out of memory and crash. So * stick with something reasonable here. */ -static INLINE int +static inline int gallivm_get_shader_param(enum pipe_shader_cap param) { switch(param) { @@ -100,7 +100,7 @@ gallivm_get_shader_param(enum pipe_shader_cap param) case PIPE_SHADER_CAP_MAX_CONTROL_FLOW_DEPTH: return LP_MAX_TGSI_NESTING; case PIPE_SHADER_CAP_MAX_INPUTS: - return PIPE_MAX_SHADER_INPUTS; + return 32; case PIPE_SHADER_CAP_MAX_OUTPUTS: return 32; case PIPE_SHADER_CAP_MAX_CONST_BUFFER_SIZE: @@ -132,6 +132,7 @@ gallivm_get_shader_param(enum pipe_shader_cap param) case PIPE_SHADER_CAP_TGSI_ANY_INOUT_DECL_RANGE: return 1; case PIPE_SHADER_CAP_DOUBLES: + return 1; case PIPE_SHADER_CAP_TGSI_DROUND_SUPPORTED: case PIPE_SHADER_CAP_TGSI_DFRACEXP_DLDEXP_SUPPORTED: case PIPE_SHADER_CAP_TGSI_FMA_SUPPORTED: diff --git a/src/gallium/auxiliary/gallivm/lp_bld_logic.c b/src/gallium/auxiliary/gallivm/lp_bld_logic.c index 80b53e5c3f8..19d30d0d63c 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_logic.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_logic.c @@ -395,7 +395,7 @@ lp_build_select(struct lp_build_context *bld, args[2] = mask; res = lp_build_intrinsic(builder, intrinsic, - arg_type, args, Elements(args)); + arg_type, args, Elements(args), 0); if (arg_type != bld->vec_type) { res = LLVMBuildBitCast(builder, res, bld->vec_type, ""); diff --git a/src/gallium/auxiliary/gallivm/lp_bld_sample.h b/src/gallium/auxiliary/gallivm/lp_bld_sample.h index 640b7e0d7e0..eba758da6ae 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_sample.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_sample.h @@ -371,7 +371,7 @@ struct lp_build_sample_context * We only support a few wrap modes in lp_build_sample_wrap_linear_int() at * this time. Return whether the given mode is supported by that function. */ -static INLINE boolean +static inline boolean lp_is_simple_wrap_mode(unsigned mode) { switch (mode) { @@ -384,7 +384,7 @@ lp_is_simple_wrap_mode(unsigned mode) } -static INLINE void +static inline void apply_sampler_swizzle(struct lp_build_sample_context *bld, LLVMValueRef *texel) { @@ -402,7 +402,7 @@ apply_sampler_swizzle(struct lp_build_sample_context *bld, * not really dimension as such, this indicates the amount of * "normal" texture coords subject to minification, wrapping etc. */ -static INLINE unsigned +static inline unsigned texture_dims(enum pipe_texture_target tex) { switch (tex) { @@ -424,7 +424,7 @@ texture_dims(enum pipe_texture_target tex) } } -static INLINE boolean +static inline boolean has_layer_coord(enum pipe_texture_target tex) { switch (tex) { diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.c index e391d8a4301..c4ae30461cb 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.c @@ -104,7 +104,7 @@ lp_build_tgsi_intrinsic( struct lp_build_context * base = &bld_base->base; emit_data->output[emit_data->chan] = lp_build_intrinsic( base->gallivm->builder, action->intr_name, - emit_data->dst_type, emit_data->args, emit_data->arg_count); + emit_data->dst_type, emit_data->args, emit_data->arg_count, 0); } LLVMValueRef @@ -175,13 +175,52 @@ void lp_build_fetch_args( unsigned src; for (src = 0; src < emit_data->info->num_src; src++) { emit_data->args[src] = lp_build_emit_fetch(bld_base, emit_data->inst, src, - emit_data->chan); + emit_data->src_chan); } emit_data->arg_count = emit_data->info->num_src; lp_build_action_set_dst_type(emit_data, bld_base, emit_data->inst->Instruction.Opcode); } +/** + * with doubles src and dst channels aren't 1:1. + * check the src/dst types for the opcode, + * 1. if neither is double then src == dst; + * 2. if dest is double + * - don't store to y or w + * - if src is double then src == dst. + * - else for f2d, d.xy = s.x + * - else for f2d, d.zw = s.y + * 3. if dst is single, src is double + * - map dst x,z to src xy; + * - map dst y,w to src zw; + */ +static int get_src_chan_idx(unsigned opcode, + int dst_chan_index) +{ + enum tgsi_opcode_type dtype = tgsi_opcode_infer_dst_type(opcode); + enum tgsi_opcode_type stype = tgsi_opcode_infer_src_type(opcode); + + if (dtype != TGSI_TYPE_DOUBLE && stype != TGSI_TYPE_DOUBLE) + return dst_chan_index; + if (dtype == TGSI_TYPE_DOUBLE) { + if (dst_chan_index == 1 || dst_chan_index == 3) + return -1; + if (stype == TGSI_TYPE_DOUBLE) + return dst_chan_index; + if (dst_chan_index == 0) + return 0; + if (dst_chan_index == 2) + return 1; + } else { + if (dst_chan_index == 0 || dst_chan_index == 2) + return 0; + if (dst_chan_index == 1 || dst_chan_index == 3) + return 2; + } + return -1; +} + /* XXX: COMMENT * It should be assumed that this function ignores writemasks */ @@ -197,7 +236,6 @@ lp_build_tgsi_inst_llvm( struct lp_build_emit_data emit_data; unsigned chan_index; LLVMValueRef val; - bld_base->pc++; if (bld_base->emit_debug) { @@ -240,7 +278,12 @@ lp_build_tgsi_inst_llvm( /* Emit the instructions */ if (info->output_mode == TGSI_OUTPUT_COMPONENTWISE && bld_base->soa) { TGSI_FOR_EACH_DST0_ENABLED_CHANNEL(inst, chan_index) { + int src_index = get_src_chan_idx(inst->Instruction.Opcode, chan_index); + /* ignore channels 1/3 in double dst */ + if (src_index == -1) + continue; emit_data.chan = chan_index; + emit_data.src_chan = src_index; if (!action->fetch_args) { lp_build_fetch_args(bld_base, &emit_data); } else { diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h index 967373ccdae..2ca9c6194b3 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi.h @@ -338,6 +338,7 @@ struct lp_build_tgsi_context struct lp_build_context uint_bld; struct lp_build_context int_bld; + struct lp_build_context dbl_bld; /** This array stores functions that are used to transform TGSI opcodes to * LLVM instructions. */ @@ -349,6 +350,9 @@ struct lp_build_tgsi_context struct lp_build_tgsi_action sqrt_action; + struct lp_build_tgsi_action drsq_action; + + struct lp_build_tgsi_action dsqrt_action; const struct tgsi_shader_info *info; lp_build_emit_fetch_fn emit_fetch_funcs[TGSI_FILE_COUNT]; @@ -558,13 +562,13 @@ struct lp_build_tgsi_aos_context }; -static INLINE struct lp_build_tgsi_soa_context * +static inline struct lp_build_tgsi_soa_context * lp_soa_context(struct lp_build_tgsi_context *bld_base) { return (struct lp_build_tgsi_soa_context *)bld_base; } -static INLINE struct lp_build_tgsi_aos_context * +static inline struct lp_build_tgsi_aos_context * lp_aos_context(struct lp_build_tgsi_context *bld_base) { return (struct lp_build_tgsi_aos_context *)bld_base; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c index 9cb42b237b7..0ad78b0ace2 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c @@ -894,6 +894,125 @@ const struct lp_build_tgsi_action xpd_action = { xpd_emit /* emit */ }; +/* TGSI_OPCODE_D2F */ +static void +d2f_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildFPTrunc(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->base.vec_type, ""); +} + +/* TGSI_OPCODE_D2I */ +static void +d2i_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildFPToSI(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->base.int_vec_type, ""); +} + +/* TGSI_OPCODE_D2U */ +static void +d2u_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildFPToUI(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->base.int_vec_type, ""); +} + +/* TGSI_OPCODE_F2D */ +static void +f2d_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildFPExt(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->dbl_bld.vec_type, ""); +} + +/* TGSI_OPCODE_U2D */ +static void +u2d_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildUIToFP(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->dbl_bld.vec_type, ""); +} + +/* TGSI_OPCODE_I2D */ +static void +i2d_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = + LLVMBuildSIToFP(bld_base->base.gallivm->builder, + emit_data->args[0], + bld_base->dbl_bld.vec_type, ""); +} + +/* TGSI_OPCODE_DMAD */ +static void +dmad_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + LLVMValueRef tmp; + tmp = lp_build_emit_llvm_binary(bld_base, TGSI_OPCODE_DMUL, + emit_data->args[0], + emit_data->args[1]); + emit_data->output[emit_data->chan] = lp_build_emit_llvm_binary(bld_base, + TGSI_OPCODE_DADD, tmp, emit_data->args[2]); +} + +/*.TGSI_OPCODE_DRCP.*/ +static void drcp_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + LLVMValueRef one; + one = lp_build_const_vec(bld_base->dbl_bld.gallivm, bld_base->dbl_bld.type, 1.0f); + emit_data->output[emit_data->chan] = LLVMBuildFDiv( + bld_base->base.gallivm->builder, + one, emit_data->args[0], ""); +} + +/* TGSI_OPCODE_DFRAC */ +static void dfrac_emit( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + LLVMValueRef tmp; + tmp = lp_build_floor(&bld_base->dbl_bld, + emit_data->args[0]); + emit_data->output[emit_data->chan] = LLVMBuildFSub(bld_base->base.gallivm->builder, + emit_data->args[0], tmp, ""); +} + void lp_set_default_actions(struct lp_build_tgsi_context * bld_base) { @@ -948,6 +1067,25 @@ lp_set_default_actions(struct lp_build_tgsi_context * bld_base) bld_base->op_actions[TGSI_OPCODE_MAX].emit = fmax_emit; bld_base->op_actions[TGSI_OPCODE_MIN].emit = fmin_emit; + + bld_base->op_actions[TGSI_OPCODE_DADD].emit = add_emit; + bld_base->op_actions[TGSI_OPCODE_DMAX].emit = fmax_emit; + bld_base->op_actions[TGSI_OPCODE_DMIN].emit = fmin_emit; + bld_base->op_actions[TGSI_OPCODE_DMUL].emit = mul_emit; + + bld_base->op_actions[TGSI_OPCODE_D2F].emit = d2f_emit; + bld_base->op_actions[TGSI_OPCODE_D2I].emit = d2i_emit; + bld_base->op_actions[TGSI_OPCODE_D2U].emit = d2u_emit; + + bld_base->op_actions[TGSI_OPCODE_F2D].emit = f2d_emit; + bld_base->op_actions[TGSI_OPCODE_I2D].emit = i2d_emit; + bld_base->op_actions[TGSI_OPCODE_U2D].emit = u2d_emit; + + bld_base->op_actions[TGSI_OPCODE_DMAD].emit = dmad_emit; + + bld_base->op_actions[TGSI_OPCODE_DRCP].emit = drcp_emit; + bld_base->op_actions[TGSI_OPCODE_DFRAC].emit = dfrac_emit; + } /* CPU Only default actions */ @@ -1792,6 +1930,107 @@ xor_emit_cpu( emit_data->args[1]); } +/* TGSI_OPCODE_DABS (CPU Only) */ +static void +dabs_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = lp_build_abs(&bld_base->dbl_bld, + emit_data->args[0]); +} + +/* TGSI_OPCODE_DNEG (CPU Only) */ +static void +dneg_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = lp_build_sub(&bld_base->dbl_bld, + bld_base->dbl_bld.zero, + emit_data->args[0]); +} + +/* TGSI_OPCODE_DSET Helper (CPU Only) */ +static void +dset_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data, + unsigned pipe_func) +{ + LLVMBuilderRef builder = bld_base->base.gallivm->builder; + LLVMValueRef cond = lp_build_cmp(&bld_base->dbl_bld, pipe_func, + emit_data->args[0], emit_data->args[1]); + /* arguments were 64 bit but store as 32 bit */ + cond = LLVMBuildTrunc(builder, cond, bld_base->int_bld.int_vec_type, ""); + emit_data->output[emit_data->chan] = cond; +} + +/* TGSI_OPCODE_DSEQ (CPU Only) */ +static void +dseq_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + dset_emit_cpu(action, bld_base, emit_data, PIPE_FUNC_EQUAL); +} + +/* TGSI_OPCODE_DSGE (CPU Only) */ +static void +dsge_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + dset_emit_cpu(action, bld_base, emit_data, PIPE_FUNC_GEQUAL); +} + +/* TGSI_OPCODE_DSLT (CPU Only) */ +static void +dslt_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + dset_emit_cpu(action, bld_base, emit_data, PIPE_FUNC_LESS); +} + +/* TGSI_OPCODE_DSNE (CPU Only) */ +static void +dsne_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + dset_emit_cpu(action, bld_base, emit_data, PIPE_FUNC_NOTEQUAL); +} + +/* Double Reciprocal squareroot (CPU Only) */ +static void +drecip_sqrt_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = lp_build_rsqrt(&bld_base->dbl_bld, + emit_data->args[0]); +} + +/* Double Squareroot (CPU Only) */ +static void +dsqrt_emit_cpu( + const struct lp_build_tgsi_action * action, + struct lp_build_tgsi_context * bld_base, + struct lp_build_emit_data * emit_data) +{ + emit_data->output[emit_data->chan] = lp_build_sqrt(&bld_base->dbl_bld, + emit_data->args[0]); +} + void lp_set_default_actions_cpu( struct lp_build_tgsi_context * bld_base) @@ -1864,4 +2103,14 @@ lp_set_default_actions_cpu( bld_base->op_actions[TGSI_OPCODE_XOR].emit = xor_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DABS].emit = dabs_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DNEG].emit = dneg_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DSEQ].emit = dseq_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DSGE].emit = dsge_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DSLT].emit = dslt_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DSNE].emit = dsne_emit_cpu; + + bld_base->op_actions[TGSI_OPCODE_DRSQ].emit = drecip_sqrt_emit_cpu; + bld_base->op_actions[TGSI_OPCODE_DSQRT].emit = dsqrt_emit_cpu; + } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.h b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.h index fc7fdbdd231..463d44eb450 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.h @@ -71,6 +71,11 @@ struct lp_build_emit_data { */ unsigned chan; + /** + * This is used to specify the src channel to read from for doubles. + */ + unsigned src_chan; + /** The lp_build_tgsi_action::emit 'executes' the opcode and writes the * results to this array. */ diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_info.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_info.c index 55acea83799..906a1745551 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_info.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_info.c @@ -462,7 +462,7 @@ analyse_instruction(struct analysis_context *ctx, } -static INLINE void +static inline void dump_info(const struct tgsi_token *tokens, struct lp_tgsi_info *info) { diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c index 268379e7d13..fae604e2f9c 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa.c @@ -106,7 +106,7 @@ emit_dump_reg(struct gallivm_state *gallivm, * Return the context for the current function. * (always 'main', if shader doesn't do any function calls) */ -static INLINE struct function_ctx * +static inline struct function_ctx * func_ctx(struct lp_exec_mask *mask) { assert(mask->function_stack_size > 0); @@ -120,7 +120,7 @@ func_ctx(struct lp_exec_mask *mask) * no loop inside the current function, but we were inside * a loop inside another function, from which this one was called. */ -static INLINE boolean +static inline boolean mask_has_loop(struct lp_exec_mask *mask) { int i; @@ -138,7 +138,7 @@ mask_has_loop(struct lp_exec_mask *mask) * no switch in the current function, but we were inside * a switch inside another function, from which this one was called. */ -static INLINE boolean +static inline boolean mask_has_switch(struct lp_exec_mask *mask) { int i; @@ -156,7 +156,7 @@ mask_has_switch(struct lp_exec_mask *mask) * no conditional in the current function, but we were inside * a conditional inside another function, from which this one was called. */ -static INLINE boolean +static inline boolean mask_has_cond(struct lp_exec_mask *mask) { int i; @@ -947,15 +947,20 @@ static LLVMValueRef build_gather(struct lp_build_tgsi_context *bld_base, LLVMValueRef base_ptr, LLVMValueRef indexes, - LLVMValueRef overflow_mask) + LLVMValueRef overflow_mask, + LLVMValueRef indexes2) { struct gallivm_state *gallivm = bld_base->base.gallivm; LLVMBuilderRef builder = gallivm->builder; struct lp_build_context *uint_bld = &bld_base->uint_bld; struct lp_build_context *bld = &bld_base->base; - LLVMValueRef res = bld->undef; + LLVMValueRef res; unsigned i; + if (indexes2) + res = LLVMGetUndef(LLVMVectorType(LLVMFloatTypeInContext(gallivm->context), bld_base->base.type.length * 2)); + else + res = bld->undef; /* * overflow_mask is a vector telling us which channels * in the vector overflowed. We use the overflow behavior for @@ -976,26 +981,47 @@ build_gather(struct lp_build_tgsi_context *bld_base, * control flow. */ indexes = lp_build_select(uint_bld, overflow_mask, uint_bld->zero, indexes); + if (indexes2) + indexes2 = lp_build_select(uint_bld, overflow_mask, uint_bld->zero, indexes2); } /* * Loop over elements of index_vec, load scalar value, insert it into 'res'. */ - for (i = 0; i < bld->type.length; i++) { - LLVMValueRef ii = lp_build_const_int32(bld->gallivm, i); - LLVMValueRef index = LLVMBuildExtractElement(builder, - indexes, ii, ""); + for (i = 0; i < bld->type.length * (indexes2 ? 2 : 1); i++) { + LLVMValueRef si, di; + LLVMValueRef index; LLVMValueRef scalar_ptr, scalar; + di = lp_build_const_int32(bld->gallivm, i); + if (indexes2) + si = lp_build_const_int32(bld->gallivm, i >> 1); + else + si = di; + + if (indexes2 && (i & 1)) { + index = LLVMBuildExtractElement(builder, + indexes2, si, ""); + } else { + index = LLVMBuildExtractElement(builder, + indexes, si, ""); + } scalar_ptr = LLVMBuildGEP(builder, base_ptr, &index, 1, "gather_ptr"); scalar = LLVMBuildLoad(builder, scalar_ptr, ""); - res = LLVMBuildInsertElement(builder, res, scalar, ii, ""); + res = LLVMBuildInsertElement(builder, res, scalar, di, ""); } if (overflow_mask) { - res = lp_build_select(bld, overflow_mask, bld->zero, res); + if (indexes2) { + res = LLVMBuildBitCast(builder, res, bld_base->dbl_bld.vec_type, ""); + overflow_mask = LLVMBuildSExt(builder, overflow_mask, + bld_base->dbl_bld.int_vec_type, ""); + res = lp_build_select(&bld_base->dbl_bld, overflow_mask, + bld_base->dbl_bld.zero, res); + } else + res = lp_build_select(bld, overflow_mask, bld->zero, res); } return res; @@ -1139,8 +1165,10 @@ stype_to_fetch(struct lp_build_tgsi_context * bld_base, case TGSI_TYPE_SIGNED: bld_fetch = &bld_base->int_bld; break; - case TGSI_TYPE_VOID: case TGSI_TYPE_DOUBLE: + bld_fetch = &bld_base->dbl_bld; + break; + case TGSI_TYPE_VOID: default: assert(0); bld_fetch = NULL; @@ -1216,6 +1244,7 @@ emit_fetch_constant( lp_build_const_int_vec(gallivm, uint_bld->type, swizzle); LLVMValueRef index_vec; /* index into the const buffer */ LLVMValueRef overflow_mask; + LLVMValueRef index_vec2 = NULL; indirect_index = get_indirect_index(bld, reg->Register.File, @@ -1235,22 +1264,33 @@ emit_fetch_constant( index_vec = lp_build_shl_imm(uint_bld, indirect_index, 2); index_vec = lp_build_add(uint_bld, index_vec, swizzle_vec); + if (stype == TGSI_TYPE_DOUBLE) { + LLVMValueRef swizzle_vec2; + swizzle_vec2 = lp_build_const_int_vec(gallivm, uint_bld->type, swizzle + 1); + index_vec2 = lp_build_shl_imm(uint_bld, indirect_index, 2); + index_vec2 = lp_build_add(uint_bld, index_vec2, swizzle_vec2); + } /* Gather values from the constant buffer */ - res = build_gather(bld_base, consts_ptr, index_vec, overflow_mask); + res = build_gather(bld_base, consts_ptr, index_vec, overflow_mask, index_vec2); } else { LLVMValueRef index; /* index into the const buffer */ LLVMValueRef scalar, scalar_ptr; - + struct lp_build_context *bld_broad = &bld_base->base; index = lp_build_const_int32(gallivm, reg->Register.Index * 4 + swizzle); scalar_ptr = LLVMBuildGEP(builder, consts_ptr, &index, 1, ""); + if (stype == TGSI_TYPE_DOUBLE) { + LLVMTypeRef dptr_type = LLVMPointerType(LLVMDoubleTypeInContext(gallivm->context), 0); + scalar_ptr = LLVMBuildBitCast(builder, scalar_ptr, dptr_type, ""); + bld_broad = &bld_base->dbl_bld; + } scalar = LLVMBuildLoad(builder, scalar_ptr, ""); - res = lp_build_broadcast_scalar(&bld_base->base, scalar); + res = lp_build_broadcast_scalar(bld_broad, scalar); } - if (stype == TGSI_TYPE_SIGNED || stype == TGSI_TYPE_UNSIGNED) { + if (stype == TGSI_TYPE_SIGNED || stype == TGSI_TYPE_UNSIGNED || stype == TGSI_TYPE_DOUBLE) { struct lp_build_context *bld_fetch = stype_to_fetch(bld_base, stype); res = LLVMBuildBitCast(builder, res, bld_fetch->vec_type, ""); } @@ -1258,6 +1298,39 @@ emit_fetch_constant( return res; } +/** + * Fetch double values from two separate channels. + * Doubles are stored split across two channels, like xy and zw. + * This function creates a set of 16 floats, + * extracts the values from the two channels, + * puts them in the correct place, then casts to 8 doubles. + */ +static LLVMValueRef +emit_fetch_double( + struct lp_build_tgsi_context * bld_base, + enum tgsi_opcode_type stype, + LLVMValueRef input, + LLVMValueRef input2) +{ + struct lp_build_tgsi_soa_context * bld = lp_soa_context(bld_base); + struct gallivm_state *gallivm = bld->bld_base.base.gallivm; + LLVMBuilderRef builder = gallivm->builder; + LLVMValueRef res; + struct lp_build_context *bld_fetch = stype_to_fetch(bld_base, stype); + int i; + LLVMValueRef shuffles[16]; + int len = bld_base->base.type.length * 2; + assert(len <= 16); + + for (i = 0; i < bld_base->base.type.length * 2; i+=2) { + shuffles[i] = lp_build_const_int32(gallivm, i / 2); + shuffles[i + 1] = lp_build_const_int32(gallivm, i / 2 + bld_base->base.type.length); + } + res = LLVMBuildShuffleVector(builder, input, input2, LLVMConstVector(shuffles, len), ""); + + return LLVMBuildBitCast(builder, res, bld_fetch->vec_type, ""); +} + static LLVMValueRef emit_fetch_immediate( struct lp_build_tgsi_context * bld_base, @@ -1281,7 +1354,7 @@ emit_fetch_immediate( if (reg->Register.Indirect) { LLVMValueRef indirect_index; LLVMValueRef index_vec; /* index into the immediate register array */ - + LLVMValueRef index_vec2 = NULL; indirect_index = get_indirect_index(bld, reg->Register.File, reg->Register.Index, @@ -1296,25 +1369,46 @@ emit_fetch_immediate( indirect_index, swizzle, FALSE); - + if (stype == TGSI_TYPE_DOUBLE) + index_vec2 = get_soa_array_offsets(&bld_base->uint_bld, + indirect_index, + swizzle + 1, + FALSE); /* Gather values from the immediate register array */ - res = build_gather(bld_base, imms_array, index_vec, NULL); + res = build_gather(bld_base, imms_array, index_vec, NULL, index_vec2); } else { LLVMValueRef lindex = lp_build_const_int32(gallivm, reg->Register.Index * 4 + swizzle); LLVMValueRef imms_ptr = LLVMBuildGEP(builder, bld->imms_array, &lindex, 1, ""); res = LLVMBuildLoad(builder, imms_ptr, ""); + + if (stype == TGSI_TYPE_DOUBLE) { + LLVMValueRef lindex1; + LLVMValueRef imms_ptr2; + LLVMValueRef res2; + + lindex1 = lp_build_const_int32(gallivm, + reg->Register.Index * 4 + swizzle + 1); + imms_ptr2 = LLVMBuildGEP(builder, + bld->imms_array, &lindex1, 1, ""); + res2 = LLVMBuildLoad(builder, imms_ptr2, ""); + res = emit_fetch_double(bld_base, stype, res, res2); + } } } else { res = bld->immediates[reg->Register.Index][swizzle]; + if (stype == TGSI_TYPE_DOUBLE) + res = emit_fetch_double(bld_base, stype, res, bld->immediates[reg->Register.Index][swizzle + 1]); } if (stype == TGSI_TYPE_UNSIGNED) { res = LLVMBuildBitCast(builder, res, bld_base->uint_bld.vec_type, ""); } else if (stype == TGSI_TYPE_SIGNED) { res = LLVMBuildBitCast(builder, res, bld_base->int_bld.vec_type, ""); + } else if (stype == TGSI_TYPE_DOUBLE) { + res = LLVMBuildBitCast(builder, res, bld_base->dbl_bld.vec_type, ""); } return res; } @@ -1334,6 +1428,7 @@ emit_fetch_input( if (reg->Register.Indirect) { LLVMValueRef indirect_index; LLVMValueRef index_vec; /* index into the input reg array */ + LLVMValueRef index_vec2 = NULL; LLVMValueRef inputs_array; LLVMTypeRef fptr_type; @@ -1346,23 +1441,43 @@ emit_fetch_input( indirect_index, swizzle, TRUE); - + if (stype == TGSI_TYPE_DOUBLE) { + index_vec2 = get_soa_array_offsets(&bld_base->uint_bld, + indirect_index, + swizzle + 1, + TRUE); + } /* cast inputs_array pointer to float* */ fptr_type = LLVMPointerType(LLVMFloatTypeInContext(gallivm->context), 0); inputs_array = LLVMBuildBitCast(builder, bld->inputs_array, fptr_type, ""); /* Gather values from the input register array */ - res = build_gather(bld_base, inputs_array, index_vec, NULL); + res = build_gather(bld_base, inputs_array, index_vec, NULL, index_vec2); } else { if (bld->indirect_files & (1 << TGSI_FILE_INPUT)) { LLVMValueRef lindex = lp_build_const_int32(gallivm, reg->Register.Index * 4 + swizzle); - LLVMValueRef input_ptr = LLVMBuildGEP(builder, - bld->inputs_array, &lindex, 1, ""); + LLVMValueRef input_ptr = LLVMBuildGEP(builder, + bld->inputs_array, &lindex, 1, ""); + res = LLVMBuildLoad(builder, input_ptr, ""); + if (stype == TGSI_TYPE_DOUBLE) { + LLVMValueRef lindex1; + LLVMValueRef input_ptr2; + LLVMValueRef res2; + + lindex1 = lp_build_const_int32(gallivm, + reg->Register.Index * 4 + swizzle + 1); + input_ptr2 = LLVMBuildGEP(builder, + bld->inputs_array, &lindex1, 1, ""); + res2 = LLVMBuildLoad(builder, input_ptr2, ""); + res = emit_fetch_double(bld_base, stype, res, res2); + } } else { res = bld->inputs[reg->Register.Index][swizzle]; + if (stype == TGSI_TYPE_DOUBLE) + res = emit_fetch_double(bld_base, stype, res, bld->inputs[reg->Register.Index][swizzle + 1]); } } @@ -1372,6 +1487,8 @@ emit_fetch_input( res = LLVMBuildBitCast(builder, res, bld_base->uint_bld.vec_type, ""); } else if (stype == TGSI_TYPE_SIGNED) { res = LLVMBuildBitCast(builder, res, bld_base->int_bld.vec_type, ""); + } else if (stype == TGSI_TYPE_DOUBLE) { + res = LLVMBuildBitCast(builder, res, bld_base->dbl_bld.vec_type, ""); } return res; @@ -1413,7 +1530,7 @@ emit_fetch_gs_input( } else { attrib_index = lp_build_const_int32(gallivm, reg->Register.Index); } - + if (reg->Dimension.Indirect) { vertex_index = get_indirect_index(bld, reg->Register.File, @@ -1436,6 +1553,8 @@ emit_fetch_gs_input( res = LLVMBuildBitCast(builder, res, bld_base->uint_bld.vec_type, ""); } else if (stype == TGSI_TYPE_SIGNED) { res = LLVMBuildBitCast(builder, res, bld_base->int_bld.vec_type, ""); + } else if (stype == TGSI_TYPE_DOUBLE) { + res = LLVMBuildBitCast(builder, res, bld_base->dbl_bld.vec_type, ""); } return res; @@ -1455,7 +1574,7 @@ emit_fetch_temporary( if (reg->Register.Indirect) { LLVMValueRef indirect_index; - LLVMValueRef index_vec; /* index into the temp reg array */ + LLVMValueRef index_vec, index_vec2 = NULL; /* index into the temp reg array */ LLVMValueRef temps_array; LLVMTypeRef fptr_type; @@ -1468,21 +1587,35 @@ emit_fetch_temporary( indirect_index, swizzle, TRUE); + if (stype == TGSI_TYPE_DOUBLE) { + index_vec2 = get_soa_array_offsets(&bld_base->uint_bld, + indirect_index, + swizzle + 1, + TRUE); + } /* cast temps_array pointer to float* */ fptr_type = LLVMPointerType(LLVMFloatTypeInContext(gallivm->context), 0); temps_array = LLVMBuildBitCast(builder, bld->temps_array, fptr_type, ""); /* Gather values from the temporary register array */ - res = build_gather(bld_base, temps_array, index_vec, NULL); + res = build_gather(bld_base, temps_array, index_vec, NULL, index_vec2); } else { LLVMValueRef temp_ptr; temp_ptr = lp_get_temp_ptr_soa(bld, reg->Register.Index, swizzle); res = LLVMBuildLoad(builder, temp_ptr, ""); + + if (stype == TGSI_TYPE_DOUBLE) { + LLVMValueRef temp_ptr2, res2; + + temp_ptr2 = lp_get_temp_ptr_soa(bld, reg->Register.Index, swizzle + 1); + res2 = LLVMBuildLoad(builder, temp_ptr2, ""); + res = emit_fetch_double(bld_base, stype, res, res2); + } } - if (stype == TGSI_TYPE_SIGNED || stype == TGSI_TYPE_UNSIGNED) { + if (stype == TGSI_TYPE_SIGNED || stype == TGSI_TYPE_UNSIGNED || stype == TGSI_TYPE_DOUBLE) { struct lp_build_context *bld_fetch = stype_to_fetch(bld_base, stype); res = LLVMBuildBitCast(builder, res, bld_fetch->vec_type, ""); } @@ -1648,6 +1781,50 @@ emit_fetch_predicate( } } +/** + * store an array of 8 doubles into two arrays of 8 floats + * i.e. + * value is d0, d1, d2, d3 etc. + * each double has high and low pieces x, y + * so gets stored into the separate channels as: + * chan_ptr = d0.x, d1.x, d2.x, d3.x + * chan_ptr2 = d0.y, d1.y, d2.y, d3.y + */ +static void +emit_store_double_chan(struct lp_build_tgsi_context *bld_base, + int dtype, + LLVMValueRef chan_ptr, LLVMValueRef chan_ptr2, + LLVMValueRef pred, + LLVMValueRef value) +{ + struct lp_build_tgsi_soa_context * bld = lp_soa_context(bld_base); + struct gallivm_state *gallivm = bld_base->base.gallivm; + LLVMBuilderRef builder = gallivm->builder; + struct lp_build_context *float_bld = &bld_base->base; + int i; + LLVMValueRef temp, temp2; + LLVMValueRef shuffles[8]; + LLVMValueRef shuffles2[8]; + + for (i = 0; i < bld_base->base.type.length; i++) { + shuffles[i] = lp_build_const_int32(gallivm, i * 2); + shuffles2[i] = lp_build_const_int32(gallivm, (i * 2) + 1); + } + + temp = LLVMBuildShuffleVector(builder, value, + LLVMGetUndef(LLVMTypeOf(value)), + LLVMConstVector(shuffles, + bld_base->base.type.length), + ""); + temp2 = LLVMBuildShuffleVector(builder, value, + LLVMGetUndef(LLVMTypeOf(value)), + LLVMConstVector(shuffles2, + bld_base->base.type.length), + ""); + + lp_exec_mask_store(&bld->exec_mask, float_bld, pred, temp, chan_ptr); + lp_exec_mask_store(&bld->exec_mask, float_bld, pred, temp2, chan_ptr2); +} /** * Register store. @@ -1683,6 +1860,11 @@ emit_store_chan( } if (reg->Register.Indirect) { + /* + * Currently the mesa/st doesn't generate indirect stores + * to doubles, it normally uses MOV to do indirect stores. + */ + assert(dtype != TGSI_TYPE_DOUBLE); indirect_index = get_indirect_index(bld, reg->Register.File, reg->Register.Index, @@ -1721,13 +1903,23 @@ emit_store_chan( else { LLVMValueRef out_ptr = lp_get_output_ptr(bld, reg->Register.Index, chan_index); - lp_exec_mask_store(&bld->exec_mask, float_bld, pred, value, out_ptr); + + if (dtype == TGSI_TYPE_DOUBLE) { + LLVMValueRef out_ptr2 = lp_get_output_ptr(bld, reg->Register.Index, + chan_index + 1); + emit_store_double_chan(bld_base, dtype, out_ptr, out_ptr2, + pred, value); + } else + lp_exec_mask_store(&bld->exec_mask, float_bld, pred, value, out_ptr); } break; case TGSI_FILE_TEMPORARY: /* Temporaries are always stored as floats */ - value = LLVMBuildBitCast(builder, value, float_bld->vec_type, ""); + if (dtype != TGSI_TYPE_DOUBLE) + value = LLVMBuildBitCast(builder, value, float_bld->vec_type, ""); + else + value = LLVMBuildBitCast(builder, value, LLVMVectorType(LLVMFloatTypeInContext(gallivm->context), bld_base->base.type.length * 2), ""); if (reg->Register.Indirect) { LLVMValueRef index_vec; /* indexes into the temp registers */ @@ -1749,7 +1941,16 @@ emit_store_chan( else { LLVMValueRef temp_ptr; temp_ptr = lp_get_temp_ptr_soa(bld, reg->Register.Index, chan_index); - lp_exec_mask_store(&bld->exec_mask, float_bld, pred, value, temp_ptr); + + if (dtype == TGSI_TYPE_DOUBLE) { + LLVMValueRef temp_ptr2 = lp_get_temp_ptr_soa(bld, + reg->Register.Index, + chan_index + 1); + emit_store_double_chan(bld_base, dtype, temp_ptr, temp_ptr2, + pred, value); + } + else + lp_exec_mask_store(&bld->exec_mask, float_bld, pred, value, temp_ptr); } break; @@ -1818,13 +2019,16 @@ emit_store( { unsigned chan_index; struct lp_build_tgsi_soa_context * bld = lp_soa_context(bld_base); - + enum tgsi_opcode_type dtype = tgsi_opcode_infer_dst_type(inst->Instruction.Opcode); if(info->num_dst) { LLVMValueRef pred[TGSI_NUM_CHANNELS]; emit_fetch_predicate( bld, inst, pred ); TGSI_FOR_EACH_DST0_ENABLED_CHANNEL( inst, chan_index ) { + + if (dtype == TGSI_TYPE_DOUBLE && (chan_index == 1 || chan_index == 3)) + continue; emit_store_chan(bld_base, inst, 0, chan_index, pred[chan_index], dst[chan_index]); } } @@ -2823,6 +3027,7 @@ void lp_emit_immediate_soa( lp_build_const_vec(gallivm, bld_base->base.type, imm->u[i].Float); break; + case TGSI_IMM_FLOAT64: case TGSI_IMM_UINT32: for( i = 0; i < size; ++i ) { LLVMValueRef tmp = lp_build_const_vec(gallivm, bld_base->uint_bld.type, imm->u[i].Uint); @@ -2857,8 +3062,7 @@ void lp_emit_immediate_soa( } else { /* simply copy the immediate values into the next immediates[] slot */ unsigned i; - const uint size = imm->Immediate.NrTokens - 1; - assert(size <= 4); + assert(imm->Immediate.NrTokens - 1 <= 4); assert(bld->num_immediates < LP_MAX_INLINED_IMMEDIATES); for(i = 0; i < 4; ++i ) @@ -3674,6 +3878,12 @@ lp_build_tgsi_soa(struct gallivm_state *gallivm, lp_build_context_init(&bld.bld_base.uint_bld, gallivm, lp_uint_type(type)); lp_build_context_init(&bld.bld_base.int_bld, gallivm, lp_int_type(type)); lp_build_context_init(&bld.elem_bld, gallivm, lp_elem_type(type)); + { + struct lp_type dbl_type; + dbl_type = type; + dbl_type.width *= 2; + lp_build_context_init(&bld.bld_base.dbl_bld, gallivm, dbl_type); + } bld.mask = mask; bld.inputs = inputs; bld.outputs = outputs; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_type.h b/src/gallium/auxiliary/gallivm/lp_bld_type.h index 191cf92d2d1..7fb449fd03f 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_type.h +++ b/src/gallium/auxiliary/gallivm/lp_bld_type.h @@ -173,7 +173,7 @@ struct lp_build_context * * e.g. With PIPE_FORMAT_R32G32B32A32_FLOAT returns an lp_type with float[4] */ -static INLINE void +static inline void lp_type_from_format_desc(struct lp_type* type, const struct util_format_description *format_desc) { assert(format_desc->is_array); @@ -189,14 +189,14 @@ lp_type_from_format_desc(struct lp_type* type, const struct util_format_descript } -static INLINE void +static inline void lp_type_from_format(struct lp_type* type, enum pipe_format format) { lp_type_from_format_desc(type, util_format_description(format)); } -static INLINE unsigned +static inline unsigned lp_type_width(struct lp_type type) { return type.width * type.length; @@ -204,7 +204,7 @@ lp_type_width(struct lp_type type) /** Create scalar float type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_float(unsigned width) { struct lp_type res_type; @@ -220,7 +220,7 @@ lp_type_float(unsigned width) /** Create vector of float type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_float_vec(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -236,7 +236,7 @@ lp_type_float_vec(unsigned width, unsigned total_width) /** Create scalar int type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_int(unsigned width) { struct lp_type res_type; @@ -251,7 +251,7 @@ lp_type_int(unsigned width) /** Create vector int type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_int_vec(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -266,7 +266,7 @@ lp_type_int_vec(unsigned width, unsigned total_width) /** Create scalar uint type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_uint(unsigned width) { struct lp_type res_type; @@ -280,7 +280,7 @@ lp_type_uint(unsigned width) /** Create vector uint type */ -static INLINE struct lp_type +static inline struct lp_type lp_type_uint_vec(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -293,7 +293,7 @@ lp_type_uint_vec(unsigned width, unsigned total_width) } -static INLINE struct lp_type +static inline struct lp_type lp_type_unorm(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -307,7 +307,7 @@ lp_type_unorm(unsigned width, unsigned total_width) } -static INLINE struct lp_type +static inline struct lp_type lp_type_fixed(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -322,7 +322,7 @@ lp_type_fixed(unsigned width, unsigned total_width) } -static INLINE struct lp_type +static inline struct lp_type lp_type_ufixed(unsigned width, unsigned total_width) { struct lp_type res_type; @@ -364,7 +364,7 @@ LLVMTypeRef lp_build_int_vec_type(struct gallivm_state *gallivm, struct lp_type type); -static INLINE struct lp_type +static inline struct lp_type lp_float32_vec4_type(void) { struct lp_type type; @@ -380,7 +380,7 @@ lp_float32_vec4_type(void) } -static INLINE struct lp_type +static inline struct lp_type lp_int32_vec4_type(void) { struct lp_type type; @@ -396,7 +396,7 @@ lp_int32_vec4_type(void) } -static INLINE struct lp_type +static inline struct lp_type lp_unorm8_vec4_type(void) { struct lp_type type; diff --git a/src/gallium/auxiliary/hud/hud_context.c b/src/gallium/auxiliary/hud/hud_context.c index 6a124f7d716..95eed2698bc 100644 --- a/src/gallium/auxiliary/hud/hud_context.c +++ b/src/gallium/auxiliary/hud/hud_context.c @@ -231,18 +231,53 @@ hud_draw_string(struct hud_context *hud, unsigned x, unsigned y, } static void -number_to_human_readable(uint64_t num, boolean is_in_bytes, char *out) +number_to_human_readable(uint64_t num, uint64_t max_value, + enum pipe_driver_query_type type, char *out) { static const char *byte_units[] = - {"", " KB", " MB", " GB", " TB", " PB", " EB"}; + {" B", " KB", " MB", " GB", " TB", " PB", " EB"}; static const char *metric_units[] = {"", " k", " M", " G", " T", " P", " E"}; - const char **units = is_in_bytes ? byte_units : metric_units; - double divisor = is_in_bytes ? 1024 : 1000; - int unit = 0; + static const char *time_units[] = + {" us", " ms", " s"}; /* based on microseconds */ + static const char *hz_units[] = + {" Hz", " KHz", " MHz", " GHz"}; + static const char *percent_units[] = {"%"}; + + const char **units; + unsigned max_unit; + double divisor = (type == PIPE_DRIVER_QUERY_TYPE_BYTES) ? 1024 : 1000; + unsigned unit = 0; double d = num; - while (d > divisor) { + switch (type) { + case PIPE_DRIVER_QUERY_TYPE_MICROSECONDS: + max_unit = ARRAY_SIZE(time_units)-1; + units = time_units; + break; + case PIPE_DRIVER_QUERY_TYPE_PERCENTAGE: + max_unit = ARRAY_SIZE(percent_units)-1; + units = percent_units; + break; + case PIPE_DRIVER_QUERY_TYPE_BYTES: + max_unit = ARRAY_SIZE(byte_units)-1; + units = byte_units; + break; + case PIPE_DRIVER_QUERY_TYPE_HZ: + max_unit = ARRAY_SIZE(hz_units)-1; + units = hz_units; + break; + default: + if (max_value == 100) { + max_unit = ARRAY_SIZE(percent_units)-1; + units = percent_units; + } else { + max_unit = ARRAY_SIZE(metric_units)-1; + units = metric_units; + } + } + + while (d > divisor && unit < max_unit) { d /= divisor; unit++; } @@ -300,9 +335,9 @@ hud_pane_accumulate_vertices(struct hud_context *hud, unsigned y = pane->inner_y1 + pane->inner_height * (5 - i) / 5 - hud->font.glyph_height / 2; - number_to_human_readable(pane->max_value * i / 5, - pane->uses_byte_units, str); - hud_draw_string(hud, x, y, str); + number_to_human_readable(pane->max_value * i / 5, pane->max_value, + pane->type, str); + hud_draw_string(hud, x, y, "%s", str); } /* draw info below the pane */ @@ -311,8 +346,8 @@ hud_pane_accumulate_vertices(struct hud_context *hud, unsigned x = pane->x1 + 2; unsigned y = pane->y2 + 2 + i*hud->font.glyph_height; - number_to_human_readable(gr->current_value, - pane->uses_byte_units, str); + number_to_human_readable(gr->current_value, pane->max_value, + pane->type, str); hud_draw_string(hud, x, y, " %s: %s", gr->name, str); i++; } @@ -417,8 +452,8 @@ hud_draw(struct hud_context *hud, struct pipe_resource *tex) cso_save_blend(cso); cso_save_depth_stencil_alpha(cso); cso_save_fragment_shader(cso); - cso_save_sampler_views(cso, PIPE_SHADER_FRAGMENT); - cso_save_samplers(cso, PIPE_SHADER_FRAGMENT); + cso_save_fragment_sampler_views(cso); + cso_save_fragment_samplers(cso); cso_save_rasterizer(cso); cso_save_viewport(cso); cso_save_stream_outputs(cso); @@ -547,8 +582,8 @@ hud_draw(struct hud_context *hud, struct pipe_resource *tex) cso_restore_blend(cso); cso_restore_depth_stencil_alpha(cso); cso_restore_fragment_shader(cso); - cso_restore_sampler_views(cso, PIPE_SHADER_FRAGMENT); - cso_restore_samplers(cso, PIPE_SHADER_FRAGMENT); + cso_restore_fragment_sampler_views(cso); + cso_restore_fragment_samplers(cso); cso_restore_rasterizer(cso); cso_restore_viewport(cso); cso_restore_stream_outputs(cso); @@ -869,12 +904,16 @@ hud_parse_env_var(struct hud_context *hud, const char *env) else if (strcmp(name, "samples-passed") == 0 && has_occlusion_query(hud->pipe->screen)) { hud_pipe_query_install(pane, hud->pipe, "samples-passed", - PIPE_QUERY_OCCLUSION_COUNTER, 0, 0, FALSE); + PIPE_QUERY_OCCLUSION_COUNTER, 0, 0, + PIPE_DRIVER_QUERY_TYPE_UINT64, + PIPE_DRIVER_QUERY_RESULT_TYPE_AVERAGE); } else if (strcmp(name, "primitives-generated") == 0 && has_streamout(hud->pipe->screen)) { hud_pipe_query_install(pane, hud->pipe, "primitives-generated", - PIPE_QUERY_PRIMITIVES_GENERATED, 0, 0, FALSE); + PIPE_QUERY_PRIMITIVES_GENERATED, 0, 0, + PIPE_DRIVER_QUERY_TYPE_UINT64, + PIPE_DRIVER_QUERY_RESULT_TYPE_AVERAGE); } else { boolean processed = FALSE; @@ -901,7 +940,8 @@ hud_parse_env_var(struct hud_context *hud, const char *env) if (i < Elements(pipeline_statistics_names)) { hud_pipe_query_install(pane, hud->pipe, name, PIPE_QUERY_PIPELINE_STATISTICS, i, - 0, FALSE); + 0, PIPE_DRIVER_QUERY_TYPE_UINT64, + PIPE_DRIVER_QUERY_RESULT_TYPE_AVERAGE); processed = TRUE; } } diff --git a/src/gallium/auxiliary/hud/hud_driver_query.c b/src/gallium/auxiliary/hud/hud_driver_query.c index 603aba7e8cd..f14305ea835 100644 --- a/src/gallium/auxiliary/hud/hud_driver_query.c +++ b/src/gallium/auxiliary/hud/hud_driver_query.c @@ -43,6 +43,7 @@ struct query_info { struct pipe_context *pipe; unsigned query_type; unsigned result_index; /* unit depends on query_type */ + enum pipe_driver_query_result_type result_type; /* Ring of queries. If a query is busy, we use another slot. */ struct pipe_query *query[NUM_QUERIES]; @@ -62,7 +63,8 @@ query_new_value(struct hud_graph *gr) uint64_t now = os_time_get(); if (info->last_time) { - pipe->end_query(pipe, info->query[info->head]); + if (info->query[info->head]) + pipe->end_query(pipe, info->query[info->head]); /* read query results */ while (1) { @@ -70,7 +72,7 @@ query_new_value(struct hud_graph *gr) union pipe_query_result result; uint64_t *res64 = (uint64_t *)&result; - if (pipe->get_query_result(pipe, query, FALSE, &result)) { + if (query && pipe->get_query_result(pipe, query, FALSE, &result)) { info->results_cumulative += res64[info->result_index]; info->num_results++; @@ -88,7 +90,8 @@ query_new_value(struct hud_graph *gr) "gallium_hud: all queries are busy after %i frames, " "can't add another query\n", NUM_QUERIES); - pipe->destroy_query(pipe, info->query[info->head]); + if (info->query[info->head]) + pipe->destroy_query(pipe, info->query[info->head]); info->query[info->head] = pipe->create_query(pipe, info->query_type, 0); } @@ -106,22 +109,33 @@ query_new_value(struct hud_graph *gr) } if (info->num_results && info->last_time + gr->pane->period <= now) { - /* compute the average value across all frames */ - hud_graph_add_value(gr, info->results_cumulative / info->num_results); + uint64_t value; + + switch (info->result_type) { + default: + case PIPE_DRIVER_QUERY_RESULT_TYPE_AVERAGE: + value = info->results_cumulative / info->num_results; + break; + case PIPE_DRIVER_QUERY_RESULT_TYPE_CUMULATIVE: + value = info->results_cumulative; + break; + } + + hud_graph_add_value(gr, value); info->last_time = now; info->results_cumulative = 0; info->num_results = 0; } - - pipe->begin_query(pipe, info->query[info->head]); } else { /* initialize */ info->last_time = now; info->query[info->head] = pipe->create_query(pipe, info->query_type, 0); - pipe->begin_query(pipe, info->query[info->head]); } + + if (info->query[info->head]) + pipe->begin_query(pipe, info->query[info->head]); } static void @@ -148,7 +162,8 @@ void hud_pipe_query_install(struct hud_pane *pane, struct pipe_context *pipe, const char *name, unsigned query_type, unsigned result_index, - uint64_t max_value, boolean uses_byte_units) + uint64_t max_value, enum pipe_driver_query_type type, + enum pipe_driver_query_result_type result_type) { struct hud_graph *gr; struct query_info *info; @@ -172,12 +187,12 @@ hud_pipe_query_install(struct hud_pane *pane, struct pipe_context *pipe, info->pipe = pipe; info->query_type = query_type; info->result_index = result_index; + info->result_type = result_type; hud_pane_add_graph(pane, gr); if (pane->max_value < max_value) hud_pane_set_max_value(pane, max_value); - if (uses_byte_units) - pane->uses_byte_units = TRUE; + pane->type = type; } boolean @@ -187,7 +202,6 @@ hud_driver_query_install(struct hud_pane *pane, struct pipe_context *pipe, struct pipe_screen *screen = pipe->screen; struct pipe_driver_query_info query; unsigned num_queries, i; - boolean uses_byte_units; boolean found = FALSE; if (!screen->get_driver_query_info) @@ -206,9 +220,8 @@ hud_driver_query_install(struct hud_pane *pane, struct pipe_context *pipe, if (!found) return FALSE; - uses_byte_units = query.type == PIPE_DRIVER_QUERY_TYPE_BYTES; hud_pipe_query_install(pane, pipe, query.name, query.query_type, 0, - query.max_value.u64, uses_byte_units); + query.max_value.u64, query.type, query.result_type); return TRUE; } diff --git a/src/gallium/auxiliary/hud/hud_private.h b/src/gallium/auxiliary/hud/hud_private.h index 632926b87f5..01caf7b8b2c 100644 --- a/src/gallium/auxiliary/hud/hud_private.h +++ b/src/gallium/auxiliary/hud/hud_private.h @@ -66,7 +66,7 @@ struct hud_pane { uint64_t ceiling; unsigned dyn_ceil_last_ran; boolean dyn_ceiling; - boolean uses_byte_units; + enum pipe_driver_query_type type; uint64_t period; /* in microseconds */ struct list_head graph_list; @@ -89,7 +89,9 @@ void hud_cpu_graph_install(struct hud_pane *pane, unsigned cpu_index); void hud_pipe_query_install(struct hud_pane *pane, struct pipe_context *pipe, const char *name, unsigned query_type, unsigned result_index, - uint64_t max_value, boolean uses_byte_units); + uint64_t max_value, + enum pipe_driver_query_type type, + enum pipe_driver_query_result_type result_type); boolean hud_driver_query_install(struct hud_pane *pane, struct pipe_context *pipe, const char *name); diff --git a/src/gallium/auxiliary/nir/tgsi_to_nir.c b/src/gallium/auxiliary/nir/tgsi_to_nir.c index 061f39ac6f3..93dfb803389 100644 --- a/src/gallium/auxiliary/nir/tgsi_to_nir.c +++ b/src/gallium/auxiliary/nir/tgsi_to_nir.c @@ -184,7 +184,8 @@ ttn_emit_declaration(struct ttn_compile *c) c->samp_types[decl->Range.First + i] = type; } } else { - nir_variable *var; + bool is_array = (array_size > 1); + assert(file == TGSI_FILE_INPUT || file == TGSI_FILE_OUTPUT || file == TGSI_FILE_CONSTANT); @@ -193,76 +194,99 @@ ttn_emit_declaration(struct ttn_compile *c) if ((file == TGSI_FILE_CONSTANT) && decl->Declaration.Dimension) return; - var = rzalloc(b->shader, nir_variable); - var->data.driver_location = decl->Range.First; + if ((file == TGSI_FILE_INPUT) || (file == TGSI_FILE_OUTPUT)) { + is_array = (is_array && decl->Declaration.Array && + (decl->Array.ArrayID != 0)); + } - var->type = glsl_vec4_type(); - if (array_size > 1) - var->type = glsl_array_type(var->type, array_size); + for (i = 0; i < array_size; i++) { + unsigned idx = decl->Range.First + i; + nir_variable *var = rzalloc(b->shader, nir_variable); - switch (file) { - case TGSI_FILE_INPUT: - var->data.read_only = true; - var->data.mode = nir_var_shader_in; - var->name = ralloc_asprintf(var, "in_%d", decl->Range.First); + var->data.driver_location = idx; - /* We should probably translate to a VERT_ATTRIB_* or VARYING_SLOT_* - * instead, but nothing in NIR core is looking at the value - * currently, and this is less change to drivers. - */ - var->data.location = decl->Semantic.Name; - var->data.index = decl->Semantic.Index; + var->type = glsl_vec4_type(); + if (is_array) + var->type = glsl_array_type(var->type, array_size); - /* We definitely need to translate the interpolation field, because - * nir_print will decode it. - */ - switch (decl->Interp.Interpolate) { - case TGSI_INTERPOLATE_CONSTANT: - var->data.interpolation = INTERP_QUALIFIER_FLAT; - break; - case TGSI_INTERPOLATE_LINEAR: - var->data.interpolation = INTERP_QUALIFIER_NOPERSPECTIVE; - break; - case TGSI_INTERPOLATE_PERSPECTIVE: - var->data.interpolation = INTERP_QUALIFIER_SMOOTH; - break; - } + switch (file) { + case TGSI_FILE_INPUT: + var->data.read_only = true; + var->data.mode = nir_var_shader_in; + var->name = ralloc_asprintf(var, "in_%d", idx); - exec_list_push_tail(&b->shader->inputs, &var->node); - break; - case TGSI_FILE_OUTPUT: { - /* Since we can't load from outputs in the IR, we make temporaries - * for the outputs and emit stores to the real outputs at the end of - * the shader. - */ - nir_register *reg = nir_local_reg_create(b->impl); - reg->num_components = 4; - if (array_size > 1) - reg->num_array_elems = array_size; - - var->data.mode = nir_var_shader_out; - var->name = ralloc_asprintf(var, "out_%d", decl->Range.First); - - var->data.location = decl->Semantic.Name; - var->data.index = decl->Semantic.Index; + /* We should probably translate to a VERT_ATTRIB_* or VARYING_SLOT_* + * instead, but nothing in NIR core is looking at the value + * currently, and this is less change to drivers. + */ + var->data.location = decl->Semantic.Name; + var->data.index = decl->Semantic.Index; - for (i = 0; i < array_size; i++) { - c->output_regs[decl->Range.First + i].offset = i; - c->output_regs[decl->Range.First + i].reg = reg; + /* We definitely need to translate the interpolation field, because + * nir_print will decode it. + */ + switch (decl->Interp.Interpolate) { + case TGSI_INTERPOLATE_CONSTANT: + var->data.interpolation = INTERP_QUALIFIER_FLAT; + break; + case TGSI_INTERPOLATE_LINEAR: + var->data.interpolation = INTERP_QUALIFIER_NOPERSPECTIVE; + break; + case TGSI_INTERPOLATE_PERSPECTIVE: + var->data.interpolation = INTERP_QUALIFIER_SMOOTH; + break; + } + + exec_list_push_tail(&b->shader->inputs, &var->node); + break; + case TGSI_FILE_OUTPUT: { + /* Since we can't load from outputs in the IR, we make temporaries + * for the outputs and emit stores to the real outputs at the end of + * the shader. + */ + nir_register *reg = nir_local_reg_create(b->impl); + reg->num_components = 4; + if (is_array) + reg->num_array_elems = array_size; + + var->data.mode = nir_var_shader_out; + var->name = ralloc_asprintf(var, "out_%d", idx); + + var->data.location = decl->Semantic.Name; + if (decl->Semantic.Name == TGSI_SEMANTIC_COLOR && + decl->Semantic.Index == 0 && + c->scan->properties[TGSI_PROPERTY_FS_COLOR0_WRITES_ALL_CBUFS]) + var->data.index = -1; + else + var->data.index = decl->Semantic.Index; + + if (is_array) { + unsigned j; + for (j = 0; j < array_size; j++) { + c->output_regs[idx + j].offset = i + j; + c->output_regs[idx + j].reg = reg; + } + } else { + c->output_regs[idx].offset = i; + c->output_regs[idx].reg = reg; + } + + exec_list_push_tail(&b->shader->outputs, &var->node); } + break; + case TGSI_FILE_CONSTANT: + var->data.mode = nir_var_uniform; + var->name = ralloc_asprintf(var, "uniform_%d", idx); - exec_list_push_tail(&b->shader->outputs, &var->node); - } - break; - case TGSI_FILE_CONSTANT: - var->data.mode = nir_var_uniform; - var->name = ralloc_asprintf(var, "uniform_%d", decl->Range.First); + exec_list_push_tail(&b->shader->uniforms, &var->node); + break; + default: + unreachable("bad declaration file"); + return; + } - exec_list_push_tail(&b->shader->uniforms, &var->node); - break; - default: - unreachable("bad declaration file"); - return; + if (is_array) + break; } } @@ -975,6 +999,9 @@ static void setup_texture_info(nir_tex_instr *instr, unsigned texture) { switch (texture) { + case TGSI_TEXTURE_BUFFER: + instr->sampler_dim = GLSL_SAMPLER_DIM_BUF; + break; case TGSI_TEXTURE_1D: instr->sampler_dim = GLSL_SAMPLER_DIM_1D; break; @@ -1068,6 +1095,11 @@ ttn_tex(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src) op = nir_texop_txb; num_srcs = 2; break; + case TGSI_OPCODE_TXB2: + op = nir_texop_txb; + num_srcs = 2; + samp = 2; + break; case TGSI_OPCODE_TXL: op = nir_texop_txl; num_srcs = 2; @@ -1078,7 +1110,12 @@ ttn_tex(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src) samp = 2; break; case TGSI_OPCODE_TXF: - op = nir_texop_txf; + if (tgsi_inst->Texture.Texture == TGSI_TEXTURE_2D_MSAA || + tgsi_inst->Texture.Texture == TGSI_TEXTURE_2D_ARRAY_MSAA) { + op = nir_texop_txf_ms; + } else { + op = nir_texop_txf; + } num_srcs = 2; break; case TGSI_OPCODE_TXD: @@ -1164,6 +1201,12 @@ ttn_tex(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src) src_number++; } + if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXB2) { + instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[1], X)); + instr->src[src_number].src_type = nir_tex_src_bias; + src_number++; + } + if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXL) { instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W)); instr->src[src_number].src_type = nir_tex_src_lod; @@ -1178,7 +1221,10 @@ ttn_tex(struct ttn_compile *c, nir_alu_dest dest, nir_ssa_def **src) if (tgsi_inst->Instruction.Opcode == TGSI_OPCODE_TXF) { instr->src[src_number].src = nir_src_for_ssa(ttn_channel(b, src[0], W)); - instr->src[src_number].src_type = nir_tex_src_lod; + if (op == nir_texop_txf_ms) + instr->src[src_number].src_type = nir_tex_src_ms_index; + else + instr->src[src_number].src_type = nir_tex_src_lod; src_number++; } @@ -1472,7 +1518,7 @@ ttn_emit_instruction(struct ttn_compile *c) return; nir_ssa_def *src[TGSI_FULL_MAX_SRC_REGISTERS]; - for (i = 0; i < TGSI_FULL_MAX_SRC_REGISTERS; i++) { + for (i = 0; i < tgsi_inst->Instruction.NumSrcRegs; i++) { src[i] = ttn_get_src(c, &tgsi_inst->Src[i]); } nir_alu_dest dest = ttn_get_dest(c, tgsi_dst); @@ -1708,9 +1754,11 @@ ttn_add_output_stores(struct ttn_compile *c) for (i = 0; i < array_len; i++) { nir_intrinsic_instr *store = nir_intrinsic_instr_create(b->shader, nir_intrinsic_store_output); + unsigned loc = var->data.driver_location + i; store->num_components = 4; - store->const_index[0] = var->data.driver_location + i; - store->src[0].reg.reg = c->output_regs[var->data.driver_location].reg; + store->const_index[0] = loc; + store->src[0].reg.reg = c->output_regs[loc].reg; + store->src[0].reg.base_offset = c->output_regs[loc].offset; nir_instr_insert_after_cf_list(b->cf_node_list, &store->instr); } } diff --git a/src/gallium/auxiliary/os/os_memory_aligned.h b/src/gallium/auxiliary/os/os_memory_aligned.h index bb15f24ade3..f7d0e3652ed 100644 --- a/src/gallium/auxiliary/os/os_memory_aligned.h +++ b/src/gallium/auxiliary/os/os_memory_aligned.h @@ -55,7 +55,7 @@ add_overflow_size_t(size_t a, size_t b, size_t *res) /** * Return memory on given byte alignment */ -static INLINE void * +static inline void * os_malloc_aligned(size_t size, size_t alignment) { char *ptr, *buf; @@ -87,7 +87,7 @@ os_malloc_aligned(size_t size, size_t alignment) /** * Free memory returned by align_malloc(). */ -static INLINE void +static inline void os_free_aligned(void *ptr) { if (ptr) { diff --git a/src/gallium/auxiliary/os/os_memory_stdc.h b/src/gallium/auxiliary/os/os_memory_stdc.h index 806e5363568..c9fde06d8ac 100644 --- a/src/gallium/auxiliary/os/os_memory_stdc.h +++ b/src/gallium/auxiliary/os/os_memory_stdc.h @@ -50,7 +50,7 @@ #if defined(HAVE_POSIX_MEMALIGN) -static INLINE void * +static inline void * os_malloc_aligned(size_t size, size_t alignment) { void *ptr; diff --git a/src/gallium/auxiliary/os/os_mman.h b/src/gallium/auxiliary/os/os_mman.h index e892610bdbd..2ae0027c1c2 100644 --- a/src/gallium/auxiliary/os/os_mman.h +++ b/src/gallium/auxiliary/os/os_mman.h @@ -58,7 +58,7 @@ extern "C" { extern void *__mmap2(void *, size_t, int, int, int, size_t); -static INLINE void *os_mmap(void *addr, size_t length, int prot, int flags, +static inline void *os_mmap(void *addr, size_t length, int prot, int flags, int fd, loff_t offset) { /* offset must be aligned to 4096 (not necessarily the page size) */ @@ -78,7 +78,7 @@ static INLINE void *os_mmap(void *addr, size_t length, int prot, int flags, # define os_mmap(addr, length, prot, flags, fd, offset) \ mmap(addr, length, prot, flags, fd, offset) -static INLINE int os_munmap(void *addr, size_t length) +static inline int os_munmap(void *addr, size_t length) { /* Copied from configure code generated by AC_SYS_LARGEFILE */ #define LARGE_OFF_T ((((off_t) 1 << 31) << 31) - 1 + \ diff --git a/src/gallium/auxiliary/os/os_thread.h b/src/gallium/auxiliary/os/os_thread.h index e9da8954885..be8adcc6cf2 100644 --- a/src/gallium/auxiliary/os/os_thread.h +++ b/src/gallium/auxiliary/os/os_thread.h @@ -54,7 +54,7 @@ typedef thrd_t pipe_thread; #define PIPE_THREAD_ROUTINE( name, param ) \ int name( void *param ) -static INLINE pipe_thread pipe_thread_create( PIPE_THREAD_ROUTINE((*routine), ), void *param ) +static inline pipe_thread pipe_thread_create( PIPE_THREAD_ROUTINE((*routine), ), void *param ) { pipe_thread thread; #ifdef HAVE_PTHREAD @@ -75,17 +75,17 @@ static INLINE pipe_thread pipe_thread_create( PIPE_THREAD_ROUTINE((*routine), ), return thread; } -static INLINE int pipe_thread_wait( pipe_thread thread ) +static inline int pipe_thread_wait( pipe_thread thread ) { return thrd_join( thread, NULL ); } -static INLINE int pipe_thread_destroy( pipe_thread thread ) +static inline int pipe_thread_destroy( pipe_thread thread ) { return thrd_detach( thread ); } -static INLINE void pipe_thread_setname( const char *name ) +static inline void pipe_thread_setname( const char *name ) { #if defined(HAVE_PTHREAD) # if defined(__GNU_LIBRARY__) && defined(__GLIBC__) && defined(__GLIBC_MINOR__) && \ @@ -145,17 +145,17 @@ typedef cnd_t pipe_condvar; typedef pthread_barrier_t pipe_barrier; -static INLINE void pipe_barrier_init(pipe_barrier *barrier, unsigned count) +static inline void pipe_barrier_init(pipe_barrier *barrier, unsigned count) { pthread_barrier_init(barrier, NULL, count); } -static INLINE void pipe_barrier_destroy(pipe_barrier *barrier) +static inline void pipe_barrier_destroy(pipe_barrier *barrier) { pthread_barrier_destroy(barrier); } -static INLINE void pipe_barrier_wait(pipe_barrier *barrier) +static inline void pipe_barrier_wait(pipe_barrier *barrier) { pthread_barrier_wait(barrier); } @@ -171,7 +171,7 @@ typedef struct { pipe_condvar condvar; } pipe_barrier; -static INLINE void pipe_barrier_init(pipe_barrier *barrier, unsigned count) +static inline void pipe_barrier_init(pipe_barrier *barrier, unsigned count) { barrier->count = count; barrier->waiters = 0; @@ -180,14 +180,14 @@ static INLINE void pipe_barrier_init(pipe_barrier *barrier, unsigned count) pipe_condvar_init(barrier->condvar); } -static INLINE void pipe_barrier_destroy(pipe_barrier *barrier) +static inline void pipe_barrier_destroy(pipe_barrier *barrier) { assert(barrier->waiters == 0); pipe_mutex_destroy(barrier->mutex); pipe_condvar_destroy(barrier->condvar); } -static INLINE void pipe_barrier_wait(pipe_barrier *barrier) +static inline void pipe_barrier_wait(pipe_barrier *barrier) { pipe_mutex_lock(barrier->mutex); @@ -225,7 +225,7 @@ typedef struct } pipe_semaphore; -static INLINE void +static inline void pipe_semaphore_init(pipe_semaphore *sema, int init_val) { pipe_mutex_init(sema->mutex); @@ -233,7 +233,7 @@ pipe_semaphore_init(pipe_semaphore *sema, int init_val) sema->counter = init_val; } -static INLINE void +static inline void pipe_semaphore_destroy(pipe_semaphore *sema) { pipe_mutex_destroy(sema->mutex); @@ -241,7 +241,7 @@ pipe_semaphore_destroy(pipe_semaphore *sema) } /** Signal/increment semaphore counter */ -static INLINE void +static inline void pipe_semaphore_signal(pipe_semaphore *sema) { pipe_mutex_lock(sema->mutex); @@ -251,7 +251,7 @@ pipe_semaphore_signal(pipe_semaphore *sema) } /** Wait for semaphore counter to be greater than zero */ -static INLINE void +static inline void pipe_semaphore_wait(pipe_semaphore *sema) { pipe_mutex_lock(sema->mutex); @@ -277,7 +277,7 @@ typedef struct { #define PIPE_TSD_INIT_MAGIC 0xff8adc98 -static INLINE void +static inline void pipe_tsd_init(pipe_tsd *tsd) { if (tss_create(&tsd->key, NULL/*free*/) != 0) { @@ -286,7 +286,7 @@ pipe_tsd_init(pipe_tsd *tsd) tsd->initMagic = PIPE_TSD_INIT_MAGIC; } -static INLINE void * +static inline void * pipe_tsd_get(pipe_tsd *tsd) { if (tsd->initMagic != (int) PIPE_TSD_INIT_MAGIC) { @@ -295,7 +295,7 @@ pipe_tsd_get(pipe_tsd *tsd) return tss_get(tsd->key); } -static INLINE void +static inline void pipe_tsd_set(pipe_tsd *tsd, void *value) { if (tsd->initMagic != (int) PIPE_TSD_INIT_MAGIC) { diff --git a/src/gallium/auxiliary/os/os_time.c b/src/gallium/auxiliary/os/os_time.c index f7e4ca49c7c..3d2e4167222 100644 --- a/src/gallium/auxiliary/os/os_time.c +++ b/src/gallium/auxiliary/os/os_time.c @@ -33,11 +33,13 @@ */ -#include "pipe/p_config.h" +#include "pipe/p_defines.h" +#include "util/u_atomic.h" #if defined(PIPE_OS_UNIX) # include <time.h> /* timeval */ # include <sys/time.h> /* timeval */ +# include <sched.h> /* sched_yield */ #elif defined(PIPE_SUBSYSTEM_WINDOWS_USER) # include <windows.h> #else @@ -92,3 +94,78 @@ os_time_sleep(int64_t usecs) } #endif + + +int64_t +os_time_get_absolute_timeout(uint64_t timeout) +{ + int64_t time, abs_timeout; + + /* Also check for the type upper bound. */ + if (timeout == PIPE_TIMEOUT_INFINITE || timeout > INT64_MAX) + return PIPE_TIMEOUT_INFINITE; + + time = os_time_get_nano(); + abs_timeout = time + (int64_t)timeout; + + /* Check for overflow. */ + if (abs_timeout < time) + return PIPE_TIMEOUT_INFINITE; + + return abs_timeout; +} + + +bool +os_wait_until_zero(volatile int *var, uint64_t timeout) +{ + if (!p_atomic_read(var)) + return true; + + if (!timeout) + return false; + + if (timeout == PIPE_TIMEOUT_INFINITE) { + while (p_atomic_read(var)) { +#if defined(PIPE_OS_UNIX) + sched_yield(); +#endif + } + return true; + } + else { + int64_t start_time = os_time_get_nano(); + int64_t end_time = start_time + timeout; + + while (p_atomic_read(var)) { + if (os_time_timeout(start_time, end_time, os_time_get_nano())) + return false; + +#if defined(PIPE_OS_UNIX) + sched_yield(); +#endif + } + return true; + } +} + + +bool +os_wait_until_zero_abs_timeout(volatile int *var, int64_t timeout) +{ + if (!p_atomic_read(var)) + return true; + + if (timeout == PIPE_TIMEOUT_INFINITE) + return os_wait_until_zero(var, PIPE_TIMEOUT_INFINITE); + + while (p_atomic_read(var)) { + if (os_time_get_nano() >= timeout) + return false; + +#if defined(PIPE_OS_UNIX) + sched_yield(); +#endif + } + return true; +} diff --git a/src/gallium/auxiliary/os/os_time.h b/src/gallium/auxiliary/os/os_time.h index 4fab03cc671..9312e028809 100644 --- a/src/gallium/auxiliary/os/os_time.h +++ b/src/gallium/auxiliary/os/os_time.h @@ -45,7 +45,7 @@ #include "pipe/p_compiler.h" -#ifdef __cplusplus +#ifdef __cplusplus extern "C" { #endif @@ -60,9 +60,10 @@ os_time_get_nano(void); /* * Get the current time in microseconds from an unknown base. */ -static INLINE int64_t -os_time_get(void) { - return os_time_get_nano() / 1000; +static inline int64_t +os_time_get(void) +{ + return os_time_get_nano() / 1000; } @@ -82,19 +83,56 @@ os_time_sleep(int64_t usecs); * * Returns true if the current time has elapsed beyond the specified interval. */ -static INLINE boolean +static inline boolean os_time_timeout(int64_t start, int64_t end, int64_t curr) { - if(start <= end) + if (start <= end) return !(start <= curr && curr < end); else return !((start <= curr) || (curr < end)); } -#ifdef __cplusplus +/** + * Convert a relative timeout in nanoseconds into an absolute timeout, + * in other words, it returns current time + timeout. + * os_time_get_nano() must be monotonic. + * PIPE_TIMEOUT_INFINITE is passed through unchanged. If the calculation + * overflows, PIPE_TIMEOUT_INFINITE is returned. + */ +int64_t +os_time_get_absolute_timeout(uint64_t timeout); + + +/** + * Wait until the variable at the given memory location is zero. + * + * \param var variable + * \param timeout timeout in ns, can be anything from 0 (no wait) to + * PIPE_TIME_INFINITE (wait forever) + * \return true if the variable is zero + */ +bool +os_wait_until_zero(volatile int *var, uint64_t timeout); + + +/** + * Wait until the variable at the given memory location is zero. + * The timeout is the absolute time when the waiting should stop. If it is + * less than or equal to the current time, it only returns the status and + * doesn't wait. PIPE_TIME_INFINITE waits forever. This requires that + * os_time_get_nano is monotonic. + * + * \param var variable + * \param timeout the time in ns when the waiting should stop + * \return true if the variable is zero + */ +bool +os_wait_until_zero_abs_timeout(volatile int *var, int64_t timeout); + +#ifdef __cplusplus } #endif diff --git a/src/gallium/auxiliary/pipe-loader/Makefile.am b/src/gallium/auxiliary/pipe-loader/Makefile.am index cb6035d85c9..8c837996539 100644 --- a/src/gallium/auxiliary/pipe-loader/Makefile.am +++ b/src/gallium/auxiliary/pipe-loader/Makefile.am @@ -1,37 +1,28 @@ include Makefile.sources +include $(top_srcdir)/src/gallium/Automake.inc -AM_CPPFLAGS = $(DEFINES) \ - $(VISIBILITY_CFLAGS) \ - -I$(top_srcdir)/include \ - -I$(top_srcdir)/src \ +# XXX: check if we need the gallium/winsys include +AM_CFLAGS = \ -I$(top_srcdir)/src/loader \ - -I$(top_srcdir)/src/gallium/include \ - -I$(top_srcdir)/src/gallium/auxiliary \ - -I$(top_srcdir)/src/gallium/winsys + -I$(top_srcdir)/src/gallium/winsys \ + $(GALLIUM_PIPE_LOADER_DEFINES) \ + $(GALLIUM_CFLAGS) \ + $(VISIBILITY_CFLAGS) noinst_LTLIBRARIES = libpipe_loader.la -noinst_LTLIBRARIES += libpipe_loader_client.la + +libpipe_loader_la_SOURCES = \ + $(COMMON_SOURCES) if HAVE_DRM_LOADER_GALLIUM -AM_CFLAGS = $(LIBDRM_CFLAGS) +AM_CFLAGS += \ + $(LIBDRM_CFLAGS) -COMMON_SOURCES += $(DRM_SOURCES) +libpipe_loader_la_SOURCES += \ + $(DRM_SOURCES) -COMMON_LIBADD = \ +libpipe_loader_la_LIBADD = \ $(top_builddir)/src/loader/libloader.la endif -libpipe_loader_la_CFLAGS = \ - $(GALLIUM_PIPE_LOADER_DEFINES) \ - $(AM_CFLAGS) $(AM_CPPFLAGS) -libpipe_loader_la_SOURCES = $(COMMON_SOURCES) -libpipe_loader_la_LIBADD = $(COMMON_LIBADD) \ - $(GALLIUM_PIPE_LOADER_LIBS) - -libpipe_loader_client_la_CFLAGS = \ - $(GALLIUM_PIPE_LOADER_CLIENT_DEFINES) \ - $(AM_CFLAGS) $(AM_CPPFLAGS) -libpipe_loader_client_la_SOURCES = $(COMMON_SOURCES) -libpipe_loader_client_la_LIBADD = $(COMMON_LIBADD) \ - $(GALLIUM_PIPE_LOADER_CLIENT_LIBS) diff --git a/src/gallium/auxiliary/pipe-loader/pipe_loader.h b/src/gallium/auxiliary/pipe-loader/pipe_loader.h index 9f43f17a6e2..9b8712666bb 100644 --- a/src/gallium/auxiliary/pipe-loader/pipe_loader.h +++ b/src/gallium/auxiliary/pipe-loader/pipe_loader.h @@ -36,10 +36,6 @@ #include "pipe/p_compiler.h" #include "state_tracker/drm_driver.h" -#ifdef HAVE_PIPE_LOADER_XLIB -#include <X11/Xlib.h> -#endif - #ifdef __cplusplus extern "C" { #endif @@ -116,21 +112,6 @@ pipe_loader_configuration(struct pipe_loader_device *dev, void pipe_loader_release(struct pipe_loader_device **devs, int ndev); -#ifdef HAVE_PIPE_LOADER_XLIB - -/** - * Initialize Xlib for an associated display. - * - * This function is platform-specific. - * - * \sa pipe_loader_probe - */ -bool -pipe_loader_sw_probe_xlib(struct pipe_loader_device **devs, Display *display); - -#endif - - #ifdef HAVE_PIPE_LOADER_DRI /** @@ -195,13 +176,9 @@ pipe_loader_drm_probe(struct pipe_loader_device **devs, int ndev); * This function is platform-specific. * * \sa pipe_loader_probe - * - * \param auth_x If true, the pipe-loader will attempt to - * authenticate with the X server. */ bool -pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd, - boolean auth_x); +pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd); #endif diff --git a/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c b/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c index ffeb29906b5..1799df7e4c5 100644 --- a/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c +++ b/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c @@ -35,12 +35,6 @@ #include <xf86drm.h> #include <unistd.h> -#ifdef HAVE_PIPE_LOADER_XCB - -#include <xcb/dri2.h> - -#endif - #include "loader.h" #include "state_tracker/drm_driver.h" #include "pipe_loader_priv.h" @@ -64,78 +58,8 @@ struct pipe_loader_drm_device { static struct pipe_loader_ops pipe_loader_drm_ops; -#ifdef HAVE_PIPE_LOADER_XCB - -static xcb_screen_t * -get_xcb_screen(xcb_screen_iterator_t iter, int screen) -{ - for (; iter.rem; --screen, xcb_screen_next(&iter)) - if (screen == 0) - return iter.data; - - return NULL; -} - -#endif - -static void -pipe_loader_drm_x_auth(int fd) -{ -#ifdef HAVE_PIPE_LOADER_XCB - /* Try authenticate with the X server to give us access to devices that X - * is running on. */ - xcb_connection_t *xcb_conn; - const xcb_setup_t *xcb_setup; - xcb_screen_iterator_t s; - xcb_dri2_connect_cookie_t connect_cookie; - xcb_dri2_connect_reply_t *connect; - drm_magic_t magic; - xcb_dri2_authenticate_cookie_t authenticate_cookie; - xcb_dri2_authenticate_reply_t *authenticate; - int screen; - - xcb_conn = xcb_connect(NULL, &screen); - - if(!xcb_conn) - return; - - xcb_setup = xcb_get_setup(xcb_conn); - - if (!xcb_setup) - goto disconnect; - - s = xcb_setup_roots_iterator(xcb_setup); - connect_cookie = xcb_dri2_connect_unchecked(xcb_conn, - get_xcb_screen(s, screen)->root, - XCB_DRI2_DRIVER_TYPE_DRI); - connect = xcb_dri2_connect_reply(xcb_conn, connect_cookie, NULL); - - if (!connect || connect->driver_name_length - + connect->device_name_length == 0) { - - goto disconnect; - } - - if (drmGetMagic(fd, &magic)) - goto disconnect; - - authenticate_cookie = xcb_dri2_authenticate_unchecked(xcb_conn, - s.data->root, - magic); - authenticate = xcb_dri2_authenticate_reply(xcb_conn, - authenticate_cookie, - NULL); - FREE(authenticate); - -disconnect: - xcb_disconnect(xcb_conn); - -#endif -} - bool -pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd, - boolean auth_x) +pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd) { struct pipe_loader_drm_device *ddev = CALLOC_STRUCT(pipe_loader_drm_device); int vendor_id, chip_id; @@ -153,9 +77,6 @@ pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd, ddev->base.ops = &pipe_loader_drm_ops; ddev->fd = fd; - if (auth_x) - pipe_loader_drm_x_auth(fd); - ddev->base.driver_name = loader_get_driver_for_fd(fd, _LOADER_GALLIUM); if (!ddev->base.driver_name) goto fail; @@ -169,34 +90,19 @@ pipe_loader_drm_probe_fd(struct pipe_loader_device **dev, int fd, } static int -open_drm_minor(int minor) -{ - char path[PATH_MAX]; - snprintf(path, sizeof(path), DRM_DEV_NAME, DRM_DIR_NAME, minor); - return open(path, O_RDWR, 0); -} - -static int open_drm_render_node_minor(int minor) { char path[PATH_MAX]; snprintf(path, sizeof(path), DRM_RENDER_NODE_DEV_NAME_FORMAT, DRM_DIR_NAME, minor); - return open(path, O_RDWR, 0); + return loader_open_device(path); } int pipe_loader_drm_probe(struct pipe_loader_device **devs, int ndev) { - int i, k, fd, num_render_node_devs; - int j = 0; + int i, j, fd; - struct { - unsigned vendor_id; - unsigned chip_id; - } render_node_devs[DRM_RENDER_NODE_MAX_NODES]; - - /* Look for render nodes first */ for (i = DRM_RENDER_NODE_MIN_MINOR, j = 0; i <= DRM_RENDER_NODE_MAX_MINOR; i++) { fd = open_drm_render_node_minor(i); @@ -204,14 +110,11 @@ pipe_loader_drm_probe(struct pipe_loader_device **devs, int ndev) if (fd < 0) continue; - if (!pipe_loader_drm_probe_fd(&dev, fd, false)) { + if (!pipe_loader_drm_probe_fd(&dev, fd)) { close(fd); continue; } - render_node_devs[j].vendor_id = dev->u.pci.vendor_id; - render_node_devs[j].chip_id = dev->u.pci.chip_id; - if (j < ndev) { devs[j] = dev; } else { @@ -221,46 +124,6 @@ pipe_loader_drm_probe(struct pipe_loader_device **devs, int ndev) j++; } - num_render_node_devs = j; - - /* Next look for drm devices. */ - for (i = 0; i < DRM_MAX_MINOR; i++) { - struct pipe_loader_device *dev; - boolean duplicate = FALSE; - fd = open_drm_minor(i); - if (fd < 0) - continue; - - if (!pipe_loader_drm_probe_fd(&dev, fd, true)) { - close(fd); - continue; - } - - /* Check to make sure we aren't already accessing this device via - * render nodes. - */ - for (k = 0; k < num_render_node_devs; k++) { - if (dev->u.pci.vendor_id == render_node_devs[k].vendor_id && - dev->u.pci.chip_id == render_node_devs[k].chip_id) { - close(fd); - dev->ops->release(&dev); - duplicate = TRUE; - break; - } - } - - if (duplicate) - continue; - - if (j < ndev) { - devs[j] = dev; - } else { - dev->ops->release(&dev); - } - - j++; - } - return j; } diff --git a/src/gallium/auxiliary/pipe-loader/pipe_loader_sw.c b/src/gallium/auxiliary/pipe-loader/pipe_loader_sw.c index 3d332645231..6794930193d 100644 --- a/src/gallium/auxiliary/pipe-loader/pipe_loader_sw.c +++ b/src/gallium/auxiliary/pipe-loader/pipe_loader_sw.c @@ -32,10 +32,6 @@ #include "sw/dri/dri_sw_winsys.h" #include "sw/null/null_sw_winsys.h" #include "sw/wrapper/wrapper_sw_winsys.h" -#ifdef HAVE_PIPE_LOADER_XLIB -/* Explicitly wrap the header to ease build without X11 headers */ -#include "sw/xlib/xlib_sw_winsys.h" -#endif #include "target-helpers/inline_sw_helper.h" #include "state_tracker/drisw_api.h" @@ -53,29 +49,6 @@ static struct sw_winsys *(*backends[])() = { null_sw_create }; -#ifdef HAVE_PIPE_LOADER_XLIB -bool -pipe_loader_sw_probe_xlib(struct pipe_loader_device **devs, Display *display) -{ - struct pipe_loader_sw_device *sdev = CALLOC_STRUCT(pipe_loader_sw_device); - - if (!sdev) - return false; - - sdev->base.type = PIPE_LOADER_DEVICE_SOFTWARE; - sdev->base.driver_name = "swrast"; - sdev->base.ops = &pipe_loader_sw_ops; - sdev->ws = xlib_create_sw_winsys(display); - if (!sdev->ws) { - FREE(sdev); - return false; - } - *devs = &sdev->base; - - return true; -} -#endif - #ifdef HAVE_PIPE_LOADER_DRI bool pipe_loader_sw_probe_dri(struct pipe_loader_device **devs, struct drisw_loader_funcs *drisw_lf) diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer.h b/src/gallium/auxiliary/pipebuffer/pb_buffer.h index 03bdce31513..ba48d461d5c 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_buffer.h +++ b/src/gallium/auxiliary/pipebuffer/pb_buffer.h @@ -158,7 +158,7 @@ struct pb_vtbl /* Accessor functions for pb->vtbl: */ -static INLINE void * +static inline void * pb_map(struct pb_buffer *buf, unsigned flags, void *flush_ctx) { @@ -170,7 +170,7 @@ pb_map(struct pb_buffer *buf, } -static INLINE void +static inline void pb_unmap(struct pb_buffer *buf) { assert(buf); @@ -181,7 +181,7 @@ pb_unmap(struct pb_buffer *buf) } -static INLINE void +static inline void pb_get_base_buffer( struct pb_buffer *buf, struct pb_buffer **base_buf, pb_size *offset ) @@ -200,7 +200,7 @@ pb_get_base_buffer( struct pb_buffer *buf, } -static INLINE enum pipe_error +static inline enum pipe_error pb_validate(struct pb_buffer *buf, struct pb_validate *vl, unsigned flags) { assert(buf); @@ -211,7 +211,7 @@ pb_validate(struct pb_buffer *buf, struct pb_validate *vl, unsigned flags) } -static INLINE void +static inline void pb_fence(struct pb_buffer *buf, struct pipe_fence_handle *fence) { assert(buf); @@ -222,7 +222,7 @@ pb_fence(struct pb_buffer *buf, struct pipe_fence_handle *fence) } -static INLINE void +static inline void pb_destroy(struct pb_buffer *buf) { assert(buf); @@ -232,7 +232,7 @@ pb_destroy(struct pb_buffer *buf) buf->vtbl->destroy(buf); } -static INLINE void +static inline void pb_reference(struct pb_buffer **dst, struct pb_buffer *src) { @@ -248,7 +248,7 @@ pb_reference(struct pb_buffer **dst, * Utility function to check whether the provided alignment is consistent with * the requested or not. */ -static INLINE boolean +static inline boolean pb_check_alignment(pb_size requested, pb_size provided) { if(!requested) @@ -265,7 +265,7 @@ pb_check_alignment(pb_size requested, pb_size provided) * Utility function to check whether the provided alignment is consistent with * the requested or not. */ -static INLINE boolean +static inline boolean pb_check_usage(unsigned requested, unsigned provided) { return (requested & provided) == requested ? TRUE : FALSE; diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c index fc81e11b972..08935b4dec7 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c +++ b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c @@ -149,7 +149,7 @@ struct fenced_buffer }; -static INLINE struct fenced_manager * +static inline struct fenced_manager * fenced_manager(struct pb_manager *mgr) { assert(mgr); @@ -157,7 +157,7 @@ fenced_manager(struct pb_manager *mgr) } -static INLINE struct fenced_buffer * +static inline struct fenced_buffer * fenced_buffer(struct pb_buffer *buf) { assert(buf); @@ -240,7 +240,7 @@ fenced_manager_dump_locked(struct fenced_manager *fenced_mgr) } -static INLINE void +static inline void fenced_buffer_destroy_locked(struct fenced_manager *fenced_mgr, struct fenced_buffer *fenced_buf) { @@ -265,7 +265,7 @@ fenced_buffer_destroy_locked(struct fenced_manager *fenced_mgr, * * Reference count should be incremented before calling this function. */ -static INLINE void +static inline void fenced_buffer_add_locked(struct fenced_manager *fenced_mgr, struct fenced_buffer *fenced_buf) { @@ -289,7 +289,7 @@ fenced_buffer_add_locked(struct fenced_manager *fenced_mgr, * * Returns TRUE if the buffer was detroyed. */ -static INLINE boolean +static inline boolean fenced_buffer_remove_locked(struct fenced_manager *fenced_mgr, struct fenced_buffer *fenced_buf) { @@ -326,7 +326,7 @@ fenced_buffer_remove_locked(struct fenced_manager *fenced_mgr, * This function will release and re-acquire the mutex, so any copy of mutable * state must be discarded after calling it. */ -static INLINE enum pipe_error +static inline enum pipe_error fenced_buffer_finish_locked(struct fenced_manager *fenced_mgr, struct fenced_buffer *fenced_buf) { @@ -550,7 +550,7 @@ fenced_buffer_destroy_gpu_storage_locked(struct fenced_buffer *fenced_buf) * This function is a shorthand around pb_manager::create_buffer for * fenced_buffer_create_gpu_storage_locked()'s benefit. */ -static INLINE boolean +static inline boolean fenced_buffer_try_create_gpu_storage_locked(struct fenced_manager *fenced_mgr, struct fenced_buffer *fenced_buf) { diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c b/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c index bf1a538bf79..b97771457d6 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c +++ b/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c @@ -49,7 +49,7 @@ struct malloc_buffer extern const struct pb_vtbl malloc_buffer_vtbl; -static INLINE struct malloc_buffer * +static inline struct malloc_buffer * malloc_buffer(struct pb_buffer *buf) { assert(buf); diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c index 62df2a6b9de..47cbaeb20ac 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c @@ -50,7 +50,7 @@ struct pb_alt_manager }; -static INLINE struct pb_alt_manager * +static inline struct pb_alt_manager * pb_alt_manager(struct pb_manager *mgr) { assert(mgr); diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c index 5023687ec04..3b35049f679 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c @@ -88,7 +88,7 @@ struct pb_cache_manager }; -static INLINE struct pb_cache_buffer * +static inline struct pb_cache_buffer * pb_cache_buffer(struct pb_buffer *buf) { assert(buf); @@ -96,7 +96,7 @@ pb_cache_buffer(struct pb_buffer *buf) } -static INLINE struct pb_cache_manager * +static inline struct pb_cache_manager * pb_cache_manager(struct pb_manager *mgr) { assert(mgr); @@ -107,7 +107,7 @@ pb_cache_manager(struct pb_manager *mgr) /** * Actually destroy the buffer. */ -static INLINE void +static inline void _pb_cache_buffer_destroy(struct pb_cache_buffer *buf) { struct pb_cache_manager *mgr = buf->mgr; @@ -235,7 +235,7 @@ pb_cache_buffer_vtbl = { }; -static INLINE int +static inline int pb_cache_is_buffer_compat(struct pb_cache_buffer *buf, pb_size size, const struct pb_desc *desc) diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c index 6236afb70d1..7ad70f293a6 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c @@ -99,7 +99,7 @@ struct pb_debug_manager }; -static INLINE struct pb_debug_buffer * +static inline struct pb_debug_buffer * pb_debug_buffer(struct pb_buffer *buf) { assert(buf); @@ -107,7 +107,7 @@ pb_debug_buffer(struct pb_buffer *buf) } -static INLINE struct pb_debug_manager * +static inline struct pb_debug_manager * pb_debug_manager(struct pb_manager *mgr) { assert(mgr); @@ -123,7 +123,7 @@ static const uint8_t random_pattern[32] = { }; -static INLINE void +static inline void fill_random_pattern(uint8_t *dst, pb_size size) { pb_size i = 0; @@ -134,7 +134,7 @@ fill_random_pattern(uint8_t *dst, pb_size size) } -static INLINE boolean +static inline boolean check_random_pattern(const uint8_t *dst, pb_size size, pb_size *min_ofs, pb_size *max_ofs) { diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c index 84eb6edda34..72099ba5850 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c @@ -65,7 +65,7 @@ struct mm_pb_manager }; -static INLINE struct mm_pb_manager * +static inline struct mm_pb_manager * mm_pb_manager(struct pb_manager *mgr) { assert(mgr); @@ -83,7 +83,7 @@ struct mm_buffer }; -static INLINE struct mm_buffer * +static inline struct mm_buffer * mm_buffer(struct pb_buffer *buf) { assert(buf); diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c index 77e642ada08..c20e2dca02d 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c @@ -70,7 +70,7 @@ struct pb_ondemand_manager extern const struct pb_vtbl pb_ondemand_buffer_vtbl; -static INLINE struct pb_ondemand_buffer * +static inline struct pb_ondemand_buffer * pb_ondemand_buffer(struct pb_buffer *buf) { assert(buf); @@ -80,7 +80,7 @@ pb_ondemand_buffer(struct pb_buffer *buf) return (struct pb_ondemand_buffer *)buf; } -static INLINE struct pb_ondemand_manager * +static inline struct pb_ondemand_manager * pb_ondemand_manager(struct pb_manager *mgr) { assert(mgr); diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c index 51525b0f97c..56a5e82ece0 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c @@ -73,7 +73,7 @@ struct pool_pb_manager }; -static INLINE struct pool_pb_manager * +static inline struct pool_pb_manager * pool_pb_manager(struct pb_manager *mgr) { assert(mgr); @@ -93,7 +93,7 @@ struct pool_buffer }; -static INLINE struct pool_buffer * +static inline struct pool_buffer * pool_buffer(struct pb_buffer *buf) { assert(buf); diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c index 6a62b4f5fdb..aadeaa087f4 100644 --- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c +++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c @@ -163,7 +163,7 @@ struct pb_slab_range_manager }; -static INLINE struct pb_slab_buffer * +static inline struct pb_slab_buffer * pb_slab_buffer(struct pb_buffer *buf) { assert(buf); @@ -171,7 +171,7 @@ pb_slab_buffer(struct pb_buffer *buf) } -static INLINE struct pb_slab_manager * +static inline struct pb_slab_manager * pb_slab_manager(struct pb_manager *mgr) { assert(mgr); @@ -179,7 +179,7 @@ pb_slab_manager(struct pb_manager *mgr) } -static INLINE struct pb_slab_range_manager * +static inline struct pb_slab_range_manager * pb_slab_range_manager(struct pb_manager *mgr) { assert(mgr); diff --git a/src/gallium/auxiliary/postprocess/pp_colors.c b/src/gallium/auxiliary/postprocess/pp_colors.c index 247e4df72a4..e6ea0102eac 100644 --- a/src/gallium/auxiliary/postprocess/pp_colors.c +++ b/src/gallium/auxiliary/postprocess/pp_colors.c @@ -37,6 +37,7 @@ pp_nocolor(struct pp_queue_t *ppq, struct pipe_resource *in, { struct pp_program *p = ppq->p; + const struct pipe_sampler_state *samplers[] = {&p->sampler_point}; pp_filter_setup_in(p, in); pp_filter_setup_out(p, out); @@ -44,8 +45,7 @@ pp_nocolor(struct pp_queue_t *ppq, struct pipe_resource *in, pp_filter_set_fb(p); pp_filter_misc_state(p); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 0, &p->sampler_point); - cso_single_sampler_done(p->cso, PIPE_SHADER_FRAGMENT); + cso_set_samplers(p->cso, PIPE_SHADER_FRAGMENT, 1, samplers); cso_set_sampler_views(p->cso, PIPE_SHADER_FRAGMENT, 1, &p->view); cso_set_vertex_shader_handle(p->cso, ppq->shaders[n][0]); diff --git a/src/gallium/auxiliary/postprocess/pp_mlaa.c b/src/gallium/auxiliary/postprocess/pp_mlaa.c index 147d14de95d..024a24895c8 100644 --- a/src/gallium/auxiliary/postprocess/pp_mlaa.c +++ b/src/gallium/auxiliary/postprocess/pp_mlaa.c @@ -141,8 +141,10 @@ pp_jimenezmlaa_run(struct pp_queue_t *ppq, struct pipe_resource *in, p->pipe->clear(p->pipe, PIPE_CLEAR_STENCIL | PIPE_CLEAR_COLOR0, &p->clear_color, 0, 0); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 0, &p->sampler_point); - cso_single_sampler_done(p->cso, PIPE_SHADER_FRAGMENT); + { + const struct pipe_sampler_state *samplers[] = {&p->sampler_point}; + cso_set_samplers(p->cso, PIPE_SHADER_FRAGMENT, 1, samplers); + } cso_set_sampler_views(p->cso, PIPE_SHADER_FRAGMENT, 1, &p->view); cso_set_vertex_shader_handle(p->cso, ppq->shaders[n][1]); /* offsetvs */ @@ -168,10 +170,11 @@ pp_jimenezmlaa_run(struct pp_queue_t *ppq, struct pipe_resource *in, pp_filter_set_clear_fb(p); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 0, &p->sampler_point); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 1, &p->sampler_point); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 2, &p->sampler); - cso_single_sampler_done(p->cso, PIPE_SHADER_FRAGMENT); + { + const struct pipe_sampler_state *samplers[] = + {&p->sampler_point, &p->sampler_point, &p->sampler}; + cso_set_samplers(p->cso, PIPE_SHADER_FRAGMENT, 3, samplers); + } arr[0] = p->view; cso_set_sampler_views(p->cso, PIPE_SHADER_FRAGMENT, 3, arr); @@ -199,9 +202,11 @@ pp_jimenezmlaa_run(struct pp_queue_t *ppq, struct pipe_resource *in, u_sampler_view_default_template(&v_tmp, in, in->format); arr[0] = p->pipe->create_sampler_view(p->pipe, in, &v_tmp); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 0, &p->sampler_point); - cso_single_sampler(p->cso, PIPE_SHADER_FRAGMENT, 1, &p->sampler_point); - cso_single_sampler_done(p->cso, PIPE_SHADER_FRAGMENT); + { + const struct pipe_sampler_state *samplers[] = + {&p->sampler_point, &p->sampler_point}; + cso_set_samplers(p->cso, PIPE_SHADER_FRAGMENT, 2, samplers); + } arr[1] = p->view; cso_set_sampler_views(p->cso, PIPE_SHADER_FRAGMENT, 2, arr); diff --git a/src/gallium/auxiliary/postprocess/pp_run.c b/src/gallium/auxiliary/postprocess/pp_run.c index e76ce854442..caa2062f4cf 100644 --- a/src/gallium/auxiliary/postprocess/pp_run.c +++ b/src/gallium/auxiliary/postprocess/pp_run.c @@ -125,8 +125,8 @@ pp_run(struct pp_queue_t *ppq, struct pipe_resource *in, cso_save_rasterizer(cso); cso_save_sample_mask(cso); cso_save_min_samples(cso); - cso_save_samplers(cso, PIPE_SHADER_FRAGMENT); - cso_save_sampler_views(cso, PIPE_SHADER_FRAGMENT); + cso_save_fragment_samplers(cso); + cso_save_fragment_sampler_views(cso); cso_save_stencil_ref(cso); cso_save_stream_outputs(cso); cso_save_vertex_elements(cso); @@ -196,8 +196,8 @@ pp_run(struct pp_queue_t *ppq, struct pipe_resource *in, cso_restore_rasterizer(cso); cso_restore_sample_mask(cso); cso_restore_min_samples(cso); - cso_restore_samplers(cso, PIPE_SHADER_FRAGMENT); - cso_restore_sampler_views(cso, PIPE_SHADER_FRAGMENT); + cso_restore_fragment_samplers(cso); + cso_restore_fragment_sampler_views(cso); cso_restore_stencil_ref(cso); cso_restore_stream_outputs(cso); cso_restore_vertex_elements(cso); diff --git a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c index f9637889187..27ee8f1242a 100644 --- a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c +++ b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c @@ -510,7 +510,7 @@ void x86_mov8_imm( struct x86_function *p, struct x86_reg dst, uint8_t imm ) /** * Immediate group 1 instructions. */ -static INLINE void +static inline void x86_group1_imm( struct x86_function *p, unsigned op, struct x86_reg dst, int imm ) { @@ -2196,7 +2196,7 @@ void x86_release_func( struct x86_function *p ) } -static INLINE x86_func +static inline x86_func voidptr_to_x86_func(void *v) { union { diff --git a/src/gallium/auxiliary/rtasm/rtasm_x86sse.h b/src/gallium/auxiliary/rtasm/rtasm_x86sse.h index 498ca824cd1..b44d917cd43 100644 --- a/src/gallium/auxiliary/rtasm/rtasm_x86sse.h +++ b/src/gallium/auxiliary/rtasm/rtasm_x86sse.h @@ -136,7 +136,7 @@ enum x86_target }; /* make this read a member of x86_function if target != host is desired */ -static INLINE enum x86_target x86_target( struct x86_function* p ) +static inline enum x86_target x86_target( struct x86_function* p ) { #ifdef PIPE_ARCH_X86 return X86_32; @@ -147,7 +147,7 @@ static INLINE enum x86_target x86_target( struct x86_function* p ) #endif } -static INLINE unsigned x86_target_caps( struct x86_function* p ) +static inline unsigned x86_target_caps( struct x86_function* p ) { return p->caps; } diff --git a/src/gallium/auxiliary/target-helpers/inline_debug_helper.h b/src/gallium/auxiliary/target-helpers/inline_debug_helper.h index 0648e596549..d353ab81e34 100644 --- a/src/gallium/auxiliary/target-helpers/inline_debug_helper.h +++ b/src/gallium/auxiliary/target-helpers/inline_debug_helper.h @@ -27,7 +27,7 @@ * TODO: Audit the following *screen_create() - all of * them should return the original screen on failuire. */ -static INLINE struct pipe_screen * +static inline struct pipe_screen * debug_screen_wrap(struct pipe_screen *screen) { #if defined(GALLIUM_RBUG) diff --git a/src/gallium/auxiliary/target-helpers/inline_drm_helper.h b/src/gallium/auxiliary/target-helpers/inline_drm_helper.h index d3c331d224d..08271a760f5 100644 --- a/src/gallium/auxiliary/target-helpers/inline_drm_helper.h +++ b/src/gallium/auxiliary/target-helpers/inline_drm_helper.h @@ -42,6 +42,7 @@ #if GALLIUM_RADEONSI #include "radeon/radeon_winsys.h" #include "radeon/drm/radeon_drm_public.h" +#include "amdgpu/drm/amdgpu_public.h" #include "radeonsi/si_public.h" #endif @@ -228,7 +229,12 @@ pipe_radeonsi_create_screen(int fd) { struct radeon_winsys *rw; - rw = radeon_drm_winsys_create(fd, radeonsi_screen_create); + /* First, try amdgpu. */ + rw = amdgpu_winsys_create(fd, radeonsi_screen_create); + + if (!rw) + rw = radeon_drm_winsys_create(fd, radeonsi_screen_create); + return rw ? debug_screen_wrap(rw->screen) : NULL; } #endif diff --git a/src/gallium/auxiliary/target-helpers/inline_sw_helper.h b/src/gallium/auxiliary/target-helpers/inline_sw_helper.h index d8cee2b2917..5f46552f6c3 100644 --- a/src/gallium/auxiliary/target-helpers/inline_sw_helper.h +++ b/src/gallium/auxiliary/target-helpers/inline_sw_helper.h @@ -20,7 +20,7 @@ #endif -static INLINE struct pipe_screen * +static inline struct pipe_screen * sw_screen_create_named(struct sw_winsys *winsys, const char *driver) { struct pipe_screen *screen = NULL; @@ -39,7 +39,7 @@ sw_screen_create_named(struct sw_winsys *winsys, const char *driver) } -static INLINE struct pipe_screen * +static inline struct pipe_screen * sw_screen_create(struct sw_winsys *winsys) { const char *default_driver; @@ -71,7 +71,7 @@ PUBLIC const __DRIextension **__driDriverGetExtensions_swrast(void) return galliumsw_driver_extensions; } -INLINE struct pipe_screen * +inline struct pipe_screen * drisw_create_screen(struct drisw_loader_funcs *lf) { struct sw_winsys *winsys = NULL; @@ -98,7 +98,7 @@ drisw_create_screen(struct drisw_loader_funcs *lf) extern struct pipe_screen *ninesw_create_screen(struct pipe_screen *screen); -INLINE struct pipe_screen * +inline struct pipe_screen * ninesw_create_screen(struct pipe_screen *pscreen) { struct sw_winsys *winsys = NULL; diff --git a/src/gallium/auxiliary/target-helpers/inline_wrapper_sw_helper.h b/src/gallium/auxiliary/target-helpers/inline_wrapper_sw_helper.h index 0a2e215352b..4f38ba9f919 100644 --- a/src/gallium/auxiliary/target-helpers/inline_wrapper_sw_helper.h +++ b/src/gallium/auxiliary/target-helpers/inline_wrapper_sw_helper.h @@ -9,7 +9,7 @@ * Try to wrap a hw screen with a software screen. * On failure will return given screen. */ -static INLINE struct pipe_screen * +static inline struct pipe_screen * sw_screen_wrap(struct pipe_screen *screen) { #if defined(GALLIUM_SOFTPIPE) || defined(GALLIUM_LLVMPIPE) diff --git a/src/gallium/auxiliary/tgsi/tgsi_dump.c b/src/gallium/auxiliary/tgsi/tgsi_dump.c index c80d7a20481..8ceb5b47584 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_dump.c +++ b/src/gallium/auxiliary/tgsi/tgsi_dump.c @@ -48,6 +48,7 @@ struct dump_ctx int indent; uint indentation; + FILE *file; void (*dump_printf)(struct dump_ctx *ctx, const char *format, ...); }; @@ -58,7 +59,10 @@ dump_ctx_printf(struct dump_ctx *ctx, const char *format, ...) va_list ap; (void)ctx; va_start(ap, format); - _debug_vprintf(format, ap); + if (ctx->file) + vfprintf(ctx->file, format, ap); + else + _debug_vprintf(format, ap); va_end(ap); } @@ -659,9 +663,7 @@ prolog( } void -tgsi_dump( - const struct tgsi_token *tokens, - uint flags ) +tgsi_dump_to_file(const struct tgsi_token *tokens, uint flags, FILE *file) { struct dump_ctx ctx; @@ -677,10 +679,17 @@ tgsi_dump( ctx.indent = 0; ctx.dump_printf = dump_ctx_printf; ctx.indentation = 0; + ctx.file = file; tgsi_iterate_shader( tokens, &ctx.iter ); } +void +tgsi_dump(const struct tgsi_token *tokens, uint flags) +{ + tgsi_dump_to_file(tokens, flags, NULL); +} + struct str_dump_ctx { struct dump_ctx base; @@ -733,6 +742,7 @@ tgsi_dump_str( ctx.base.indent = 0; ctx.base.dump_printf = &str_dump_ctx_printf; ctx.base.indentation = 0; + ctx.base.file = NULL; ctx.str = str; ctx.str[0] = 0; @@ -756,6 +766,7 @@ tgsi_dump_instruction_str( ctx.base.indent = 0; ctx.base.dump_printf = &str_dump_ctx_printf; ctx.base.indentation = 0; + ctx.base.file = NULL; ctx.str = str; ctx.str[0] = 0; diff --git a/src/gallium/auxiliary/tgsi/tgsi_dump.h b/src/gallium/auxiliary/tgsi/tgsi_dump.h index bc873a54ae9..7c8f92ee7bc 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_dump.h +++ b/src/gallium/auxiliary/tgsi/tgsi_dump.h @@ -32,6 +32,8 @@ #include "pipe/p_defines.h" #include "pipe/p_shader_tokens.h" +#include <stdio.h> + #if defined __cplusplus extern "C" { #endif @@ -44,6 +46,9 @@ tgsi_dump_str( size_t size); void +tgsi_dump_to_file(const struct tgsi_token *tokens, uint flags, FILE *file); + +void tgsi_dump( const struct tgsi_token *tokens, uint flags ); diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.c b/src/gallium/auxiliary/tgsi/tgsi_exec.c index 44000ffdb6c..75cd0d53c5a 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_exec.c +++ b/src/gallium/auxiliary/tgsi/tgsi_exec.c @@ -735,7 +735,7 @@ static const union tgsi_exec_channel M128Vec = { * not lead to crashes, etc. But when debugging, it's helpful to catch * them. */ -static INLINE void +static inline void check_inf_or_nan(const union tgsi_exec_channel *chan) { assert(!util_is_inf_or_nan((chan)->f[0])); diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.h b/src/gallium/auxiliary/tgsi/tgsi_exec.h index 208640cfd46..5d56aab2216 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_exec.h +++ b/src/gallium/auxiliary/tgsi/tgsi_exec.h @@ -213,7 +213,7 @@ struct tgsi_sampler * input register files, this is the stride between two 1D * arrays. */ -#define TGSI_EXEC_MAX_INPUT_ATTRIBS PIPE_MAX_SHADER_INPUTS +#define TGSI_EXEC_MAX_INPUT_ATTRIBS 32 /* The maximum number of bytes per constant buffer. */ @@ -386,7 +386,7 @@ boolean tgsi_check_soa_dependencies(const struct tgsi_full_instruction *inst); -static INLINE void +static inline void tgsi_set_kill_mask(struct tgsi_exec_machine *mach, unsigned mask) { mach->Temps[TGSI_EXEC_TEMP_KILMASK_I].xyzw[TGSI_EXEC_TEMP_KILMASK_C].u[0] = @@ -395,7 +395,7 @@ tgsi_set_kill_mask(struct tgsi_exec_machine *mach, unsigned mask) /** Set execution mask values prior to executing the shader */ -static INLINE void +static inline void tgsi_set_exec_mask(struct tgsi_exec_machine *mach, boolean ch0, boolean ch1, boolean ch2, boolean ch3) { @@ -414,7 +414,7 @@ tgsi_exec_set_constant_buffers(struct tgsi_exec_machine *mach, const unsigned *buf_sizes); -static INLINE int +static inline int tgsi_exec_get_shader_param(enum pipe_shader_cap param) { switch(param) { diff --git a/src/gallium/auxiliary/tgsi/tgsi_info.c b/src/gallium/auxiliary/tgsi/tgsi_info.c index 929531109e5..fb29ea0d53d 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_info.c +++ b/src/gallium/auxiliary/tgsi/tgsi_info.c @@ -316,7 +316,7 @@ tgsi_get_processor_name( uint processor ) * * MOV and UCMP is special so return VOID */ -static INLINE enum tgsi_opcode_type +static inline enum tgsi_opcode_type tgsi_opcode_infer_type( uint opcode ) { switch (opcode) { @@ -374,7 +374,34 @@ tgsi_opcode_infer_type( uint opcode ) case TGSI_OPCODE_IMUL_HI: case TGSI_OPCODE_IBFE: case TGSI_OPCODE_IMSB: + case TGSI_OPCODE_DSEQ: + case TGSI_OPCODE_DSGE: + case TGSI_OPCODE_DSLT: + case TGSI_OPCODE_DSNE: return TGSI_TYPE_SIGNED; + case TGSI_OPCODE_DADD: + case TGSI_OPCODE_DABS: + case TGSI_OPCODE_DFMA: + case TGSI_OPCODE_DNEG: + case TGSI_OPCODE_DMUL: + case TGSI_OPCODE_DMAX: + case TGSI_OPCODE_DMIN: + case TGSI_OPCODE_DRCP: + case TGSI_OPCODE_DSQRT: + case TGSI_OPCODE_DMAD: + case TGSI_OPCODE_DLDEXP: + case TGSI_OPCODE_DFRACEXP: + case TGSI_OPCODE_DFRAC: + case TGSI_OPCODE_DRSQ: + case TGSI_OPCODE_DTRUNC: + case TGSI_OPCODE_DCEIL: + case TGSI_OPCODE_DFLR: + case TGSI_OPCODE_DROUND: + case TGSI_OPCODE_DSSG: + case TGSI_OPCODE_F2D: + case TGSI_OPCODE_I2D: + case TGSI_OPCODE_U2D: + return TGSI_TYPE_DOUBLE; default: return TGSI_TYPE_FLOAT; } @@ -391,6 +418,7 @@ tgsi_opcode_infer_src_type( uint opcode ) case TGSI_OPCODE_TXF: case TGSI_OPCODE_BREAKC: case TGSI_OPCODE_U2F: + case TGSI_OPCODE_U2D: case TGSI_OPCODE_UADD: case TGSI_OPCODE_SWITCH: case TGSI_OPCODE_CASE: @@ -400,10 +428,12 @@ tgsi_opcode_infer_src_type( uint opcode ) return TGSI_TYPE_UNSIGNED; case TGSI_OPCODE_IMUL_HI: case TGSI_OPCODE_I2F: + case TGSI_OPCODE_I2D: return TGSI_TYPE_SIGNED; case TGSI_OPCODE_ARL: case TGSI_OPCODE_ARR: case TGSI_OPCODE_TXQ_LZ: + case TGSI_OPCODE_F2D: case TGSI_OPCODE_F2I: case TGSI_OPCODE_F2U: case TGSI_OPCODE_FSEQ: @@ -412,6 +442,14 @@ tgsi_opcode_infer_src_type( uint opcode ) case TGSI_OPCODE_FSNE: case TGSI_OPCODE_UCMP: return TGSI_TYPE_FLOAT; + case TGSI_OPCODE_D2F: + case TGSI_OPCODE_D2U: + case TGSI_OPCODE_D2I: + case TGSI_OPCODE_DSEQ: + case TGSI_OPCODE_DSGE: + case TGSI_OPCODE_DSLT: + case TGSI_OPCODE_DSNE: + return TGSI_TYPE_DOUBLE; default: return tgsi_opcode_infer_type(opcode); } diff --git a/src/gallium/auxiliary/tgsi/tgsi_parse.c b/src/gallium/auxiliary/tgsi/tgsi_parse.c index 1162b265522..0729b5d2426 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_parse.c +++ b/src/gallium/auxiliary/tgsi/tgsi_parse.c @@ -36,7 +36,7 @@ tgsi_parse_init( const struct tgsi_token *tokens ) { ctx->FullHeader.Header = *(struct tgsi_header *) &tokens[0]; - if( ctx->FullHeader.Header.HeaderSize >= 2 ) { + if (ctx->FullHeader.Header.HeaderSize >= 2) { ctx->FullHeader.Processor = *(struct tgsi_processor *) &tokens[1]; } else { @@ -69,7 +69,7 @@ tgsi_parse_end_of_tokens( * warnings. The warnings seem harmless on x86 but on PPC they cause * real failures. */ -static INLINE void +static inline void copy_token(void *dst, const void *src) { memcpy(dst, src, 4); @@ -113,11 +113,11 @@ tgsi_parse_token( next_token(ctx, &decl->Dim); } - if( decl->Declaration.Interpolate ) { + if (decl->Declaration.Interpolate) { next_token( ctx, &decl->Interp ); } - if( decl->Declaration.Semantic ) { + if (decl->Declaration.Semantic) { next_token( ctx, &decl->Semantic ); } @@ -129,7 +129,7 @@ tgsi_parse_token( next_token(ctx, &decl->SamplerView); } - if( decl->Declaration.Array ) { + if (decl->Declaration.Array) { next_token(ctx, &decl->Array); } @@ -190,21 +190,21 @@ tgsi_parse_token( if (inst->Instruction.Texture) { next_token( ctx, &inst->Texture); - for( i = 0; i < inst->Texture.NumOffsets; i++ ) { + for (i = 0; i < inst->Texture.NumOffsets; i++) { next_token( ctx, &inst->TexOffsets[i] ); } } assert( inst->Instruction.NumDstRegs <= TGSI_FULL_MAX_DST_REGISTERS ); - for( i = 0; i < inst->Instruction.NumDstRegs; i++ ) { + for (i = 0; i < inst->Instruction.NumDstRegs; i++) { next_token( ctx, &inst->Dst[i].Register ); - if( inst->Dst[i].Register.Indirect ) + if (inst->Dst[i].Register.Indirect) next_token( ctx, &inst->Dst[i].Indirect ); - if( inst->Dst[i].Register.Dimension ) { + if (inst->Dst[i].Register.Dimension) { next_token( ctx, &inst->Dst[i].Dimension ); /* @@ -212,21 +212,21 @@ tgsi_parse_token( */ assert( !inst->Dst[i].Dimension.Dimension ); - if( inst->Dst[i].Dimension.Indirect ) + if (inst->Dst[i].Dimension.Indirect) next_token( ctx, &inst->Dst[i].DimIndirect ); } } assert( inst->Instruction.NumSrcRegs <= TGSI_FULL_MAX_SRC_REGISTERS ); - for( i = 0; i < inst->Instruction.NumSrcRegs; i++ ) { + for (i = 0; i < inst->Instruction.NumSrcRegs; i++) { next_token( ctx, &inst->Src[i].Register ); - if( inst->Src[i].Register.Indirect ) + if (inst->Src[i].Register.Indirect) next_token( ctx, &inst->Src[i].Indirect ); - if( inst->Src[i].Register.Dimension ) { + if (inst->Src[i].Register.Dimension) { next_token( ctx, &inst->Src[i].Dimension ); /* @@ -234,7 +234,7 @@ tgsi_parse_token( */ assert( !inst->Src[i].Dimension.Dimension ); - if( inst->Src[i].Dimension.Indirect ) + if (inst->Src[i].Dimension.Indirect) next_token( ctx, &inst->Src[i].DimIndirect ); } } diff --git a/src/gallium/auxiliary/tgsi/tgsi_parse.h b/src/gallium/auxiliary/tgsi/tgsi_parse.h index cd4b2afdb8b..35e1c7cfd62 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_parse.h +++ b/src/gallium/auxiliary/tgsi/tgsi_parse.h @@ -133,7 +133,7 @@ void tgsi_parse_token( struct tgsi_parse_context *ctx ); -static INLINE unsigned +static inline unsigned tgsi_num_tokens(const struct tgsi_token *tokens) { struct tgsi_header header; diff --git a/src/gallium/auxiliary/tgsi/tgsi_sanity.c b/src/gallium/auxiliary/tgsi/tgsi_sanity.c index be4851f5dcb..d14372feb30 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_sanity.c +++ b/src/gallium/auxiliary/tgsi/tgsi_sanity.c @@ -63,7 +63,7 @@ struct sanity_check_ctx boolean print; }; -static INLINE unsigned +static inline unsigned scan_register_key(const scan_register *reg) { unsigned key = reg->file; diff --git a/src/gallium/auxiliary/tgsi/tgsi_strings.c b/src/gallium/auxiliary/tgsi/tgsi_strings.c index 6b6a14f55f5..8271ea08177 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_strings.c +++ b/src/gallium/auxiliary/tgsi/tgsi_strings.c @@ -203,7 +203,7 @@ const char *tgsi_immediate_type_names[4] = }; -static INLINE void +static inline void tgsi_strings_check(void) { STATIC_ASSERT(Elements(tgsi_semantic_names) == TGSI_SEMANTIC_COUNT); diff --git a/src/gallium/auxiliary/tgsi/tgsi_transform.h b/src/gallium/auxiliary/tgsi/tgsi_transform.h index 39d7688ab3b..ceb7c2e0f46 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_transform.h +++ b/src/gallium/auxiliary/tgsi/tgsi_transform.h @@ -94,7 +94,7 @@ struct tgsi_transform_context /** * Helper for emitting temporary register declarations. */ -static INLINE void +static inline void tgsi_transform_temp_decl(struct tgsi_transform_context *ctx, unsigned index) { @@ -108,7 +108,7 @@ tgsi_transform_temp_decl(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_input_decl(struct tgsi_transform_context *ctx, unsigned index, unsigned sem_name, unsigned sem_index, @@ -130,7 +130,7 @@ tgsi_transform_input_decl(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_sampler_decl(struct tgsi_transform_context *ctx, unsigned index) { @@ -143,7 +143,7 @@ tgsi_transform_sampler_decl(struct tgsi_transform_context *ctx, ctx->emit_declaration(ctx, &decl); } -static INLINE void +static inline void tgsi_transform_sampler_view_decl(struct tgsi_transform_context *ctx, unsigned index, unsigned target, @@ -165,7 +165,7 @@ tgsi_transform_sampler_view_decl(struct tgsi_transform_context *ctx, ctx->emit_declaration(ctx, &decl); } -static INLINE void +static inline void tgsi_transform_immediate_decl(struct tgsi_transform_context *ctx, float x, float y, float z, float w) { @@ -186,7 +186,7 @@ tgsi_transform_immediate_decl(struct tgsi_transform_context *ctx, /** * Helper for emitting 1-operand instructions. */ -static INLINE void +static inline void tgsi_transform_op1_inst(struct tgsi_transform_context *ctx, unsigned opcode, unsigned dst_file, @@ -211,7 +211,7 @@ tgsi_transform_op1_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_op2_inst(struct tgsi_transform_context *ctx, unsigned opcode, unsigned dst_file, @@ -240,7 +240,7 @@ tgsi_transform_op2_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_op1_swz_inst(struct tgsi_transform_context *ctx, unsigned opcode, unsigned dst_file, @@ -282,7 +282,7 @@ tgsi_transform_op1_swz_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_op2_swz_inst(struct tgsi_transform_context *ctx, unsigned opcode, unsigned dst_file, @@ -333,7 +333,7 @@ tgsi_transform_op2_swz_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_op3_swz_inst(struct tgsi_transform_context *ctx, unsigned opcode, unsigned dst_file, @@ -395,7 +395,7 @@ tgsi_transform_op3_swz_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_kill_inst(struct tgsi_transform_context *ctx, unsigned src_file, unsigned src_index, @@ -419,7 +419,7 @@ tgsi_transform_kill_inst(struct tgsi_transform_context *ctx, } -static INLINE void +static inline void tgsi_transform_tex_2d_inst(struct tgsi_transform_context *ctx, unsigned dst_file, unsigned dst_index, diff --git a/src/gallium/auxiliary/tgsi/tgsi_ureg.c b/src/gallium/auxiliary/tgsi/tgsi_ureg.c index 201a849ef95..3d213195090 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_ureg.c +++ b/src/gallium/auxiliary/tgsi/tgsi_ureg.c @@ -1830,7 +1830,7 @@ void ureg_free_tokens( const struct tgsi_token *tokens ) } -static INLINE unsigned +static inline unsigned pipe_shader_from_tgsi_processor(unsigned processor) { switch (processor) { diff --git a/src/gallium/auxiliary/tgsi/tgsi_ureg.h b/src/gallium/auxiliary/tgsi/tgsi_ureg.h index 1891b068774..0aae550d60a 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_ureg.h +++ b/src/gallium/auxiliary/tgsi/tgsi_ureg.h @@ -140,7 +140,7 @@ ureg_destroy( struct ureg_program * ); /*********************************************************************** * Convenience routine: */ -static INLINE void * +static inline void * ureg_create_shader_with_so_and_destroy( struct ureg_program *p, struct pipe_context *pipe, const struct pipe_stream_output_info *so ) @@ -150,7 +150,7 @@ ureg_create_shader_with_so_and_destroy( struct ureg_program *p, return result; } -static INLINE void * +static inline void * ureg_create_shader_and_destroy( struct ureg_program *p, struct pipe_context *pipe ) { @@ -180,7 +180,7 @@ ureg_DECL_fs_input_cyl_centroid(struct ureg_program *, unsigned array_id, unsigned array_size); -static INLINE struct ureg_src +static inline struct ureg_src ureg_DECL_fs_input_cyl(struct ureg_program *ureg, unsigned semantic_name, unsigned semantic_index, @@ -195,7 +195,7 @@ ureg_DECL_fs_input_cyl(struct ureg_program *ureg, 0, 0, 1); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_DECL_fs_input(struct ureg_program *ureg, unsigned semantic_name, unsigned semantic_index, @@ -328,7 +328,7 @@ ureg_DECL_sampler_view(struct ureg_program *, unsigned return_type_w ); -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm4f( struct ureg_program *ureg, float a, float b, float c, float d) @@ -341,7 +341,7 @@ ureg_imm4f( struct ureg_program *ureg, return ureg_DECL_immediate( ureg, v, 4 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm3f( struct ureg_program *ureg, float a, float b, float c) @@ -353,7 +353,7 @@ ureg_imm3f( struct ureg_program *ureg, return ureg_DECL_immediate( ureg, v, 3 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm2f( struct ureg_program *ureg, float a, float b) { @@ -363,7 +363,7 @@ ureg_imm2f( struct ureg_program *ureg, return ureg_DECL_immediate( ureg, v, 2 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm1f( struct ureg_program *ureg, float a) { @@ -372,7 +372,7 @@ ureg_imm1f( struct ureg_program *ureg, return ureg_DECL_immediate( ureg, v, 1 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm4u( struct ureg_program *ureg, unsigned a, unsigned b, unsigned c, unsigned d) @@ -385,7 +385,7 @@ ureg_imm4u( struct ureg_program *ureg, return ureg_DECL_immediate_uint( ureg, v, 4 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm3u( struct ureg_program *ureg, unsigned a, unsigned b, unsigned c) @@ -397,7 +397,7 @@ ureg_imm3u( struct ureg_program *ureg, return ureg_DECL_immediate_uint( ureg, v, 3 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm2u( struct ureg_program *ureg, unsigned a, unsigned b) { @@ -407,14 +407,14 @@ ureg_imm2u( struct ureg_program *ureg, return ureg_DECL_immediate_uint( ureg, v, 2 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm1u( struct ureg_program *ureg, unsigned a) { return ureg_DECL_immediate_uint( ureg, &a, 1 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm4i( struct ureg_program *ureg, int a, int b, int c, int d) @@ -427,7 +427,7 @@ ureg_imm4i( struct ureg_program *ureg, return ureg_DECL_immediate_int( ureg, v, 4 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm3i( struct ureg_program *ureg, int a, int b, int c) @@ -439,7 +439,7 @@ ureg_imm3i( struct ureg_program *ureg, return ureg_DECL_immediate_int( ureg, v, 3 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm2i( struct ureg_program *ureg, int a, int b) { @@ -449,7 +449,7 @@ ureg_imm2i( struct ureg_program *ureg, return ureg_DECL_immediate_int( ureg, v, 2 ); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_imm1i( struct ureg_program *ureg, int a) { @@ -459,7 +459,7 @@ ureg_imm1i( struct ureg_program *ureg, /* Where the destination register has a valid file, but an empty * writemask. */ -static INLINE boolean +static inline boolean ureg_dst_is_empty( struct ureg_dst dst ) { return dst.File != TGSI_FILE_NULL && @@ -573,7 +573,7 @@ ureg_fixup_insn_size(struct ureg_program *ureg, #define OP00( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg ) \ +static inline void ureg_##op( struct ureg_program *ureg ) \ { \ unsigned opcode = TGSI_OPCODE_##op; \ struct ureg_emit_insn_result insn; \ @@ -592,7 +592,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg ) \ } #define OP01( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_src src ) \ { \ unsigned opcode = TGSI_OPCODE_##op; \ @@ -613,7 +613,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP00_LBL( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ unsigned *label_token ) \ { \ unsigned opcode = TGSI_OPCODE_##op; \ @@ -634,7 +634,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP01_LBL( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_src src, \ unsigned *label_token ) \ { \ @@ -657,7 +657,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP10( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst ) \ { \ unsigned opcode = TGSI_OPCODE_##op; \ @@ -681,7 +681,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ #define OP11( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src ) \ { \ @@ -706,7 +706,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP12( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1 ) \ @@ -733,7 +733,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP12_TEX( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ unsigned target, \ struct ureg_src src0, \ @@ -762,7 +762,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP12_SAMPLE( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1 ) \ @@ -791,7 +791,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP13( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -820,7 +820,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP13_SAMPLE( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -851,7 +851,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP14_TEX( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ unsigned target, \ struct ureg_src src0, \ @@ -884,7 +884,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP14_SAMPLE( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -918,7 +918,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ #define OP14( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -950,7 +950,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ #define OP15( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -983,7 +983,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ } #define OP15_SAMPLE( op ) \ -static INLINE void ureg_##op( struct ureg_program *ureg, \ +static inline void ureg_##op( struct ureg_program *ureg, \ struct ureg_dst dst, \ struct ureg_src src0, \ struct ureg_src src1, \ @@ -1026,7 +1026,7 @@ static INLINE void ureg_##op( struct ureg_program *ureg, \ /*********************************************************************** * Inline helpers for manipulating register structs: */ -static INLINE struct ureg_src +static inline struct ureg_src ureg_negate( struct ureg_src reg ) { assert(reg.File != TGSI_FILE_NULL); @@ -1034,7 +1034,7 @@ ureg_negate( struct ureg_src reg ) return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_abs( struct ureg_src reg ) { assert(reg.File != TGSI_FILE_NULL); @@ -1043,7 +1043,7 @@ ureg_abs( struct ureg_src reg ) return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_swizzle( struct ureg_src reg, int x, int y, int z, int w ) { @@ -1065,13 +1065,13 @@ ureg_swizzle( struct ureg_src reg, return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_scalar( struct ureg_src reg, int x ) { return ureg_swizzle(reg, x, x, x, x); } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_writemask( struct ureg_dst reg, unsigned writemask ) { @@ -1080,7 +1080,7 @@ ureg_writemask( struct ureg_dst reg, return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_saturate( struct ureg_dst reg ) { assert(reg.File != TGSI_FILE_NULL); @@ -1088,7 +1088,7 @@ ureg_saturate( struct ureg_dst reg ) return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_predicate(struct ureg_dst reg, boolean negate, unsigned swizzle_x, @@ -1106,7 +1106,7 @@ ureg_predicate(struct ureg_dst reg, return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_indirect( struct ureg_dst reg, struct ureg_src addr ) { assert(reg.File != TGSI_FILE_NULL); @@ -1118,7 +1118,7 @@ ureg_dst_indirect( struct ureg_dst reg, struct ureg_src addr ) return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_indirect( struct ureg_src reg, struct ureg_src addr ) { assert(reg.File != TGSI_FILE_NULL); @@ -1130,7 +1130,7 @@ ureg_src_indirect( struct ureg_src reg, struct ureg_src addr ) return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_dimension( struct ureg_dst reg, int index ) { assert(reg.File != TGSI_FILE_NULL); @@ -1140,7 +1140,7 @@ ureg_dst_dimension( struct ureg_dst reg, int index ) return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_dimension( struct ureg_src reg, int index ) { assert(reg.File != TGSI_FILE_NULL); @@ -1150,7 +1150,7 @@ ureg_src_dimension( struct ureg_src reg, int index ) return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_dimension_indirect( struct ureg_dst reg, struct ureg_src addr, int index ) { @@ -1164,7 +1164,7 @@ ureg_dst_dimension_indirect( struct ureg_dst reg, struct ureg_src addr, return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_dimension_indirect( struct ureg_src reg, struct ureg_src addr, int index ) { @@ -1178,21 +1178,21 @@ ureg_src_dimension_indirect( struct ureg_src reg, struct ureg_src addr, return reg; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_array_offset(struct ureg_src reg, int offset) { reg.Index += offset; return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_array_offset( struct ureg_dst reg, int offset ) { reg.Index += offset; return reg; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_array_register(unsigned file, unsigned index, unsigned array_id) @@ -1224,14 +1224,14 @@ ureg_dst_array_register(unsigned file, return dst; } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_register(unsigned file, unsigned index) { return ureg_dst_array_register(file, index, 0); } -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst( struct ureg_src src ) { struct ureg_dst dst; @@ -1265,7 +1265,7 @@ ureg_dst( struct ureg_src src ) return dst; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_array_register(unsigned file, unsigned index, unsigned array_id) @@ -1295,14 +1295,14 @@ ureg_src_array_register(unsigned file, return src; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_register(unsigned file, unsigned index) { return ureg_src_array_register(file, index, 0); } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src( struct ureg_dst dst ) { struct ureg_src src; @@ -1332,7 +1332,7 @@ ureg_src( struct ureg_dst dst ) -static INLINE struct ureg_dst +static inline struct ureg_dst ureg_dst_undef( void ) { struct ureg_dst dst; @@ -1362,7 +1362,7 @@ ureg_dst_undef( void ) return dst; } -static INLINE struct ureg_src +static inline struct ureg_src ureg_src_undef( void ) { struct ureg_src src; @@ -1390,13 +1390,13 @@ ureg_src_undef( void ) return src; } -static INLINE boolean +static inline boolean ureg_src_is_undef( struct ureg_src src ) { return src.File == TGSI_FILE_NULL; } -static INLINE boolean +static inline boolean ureg_dst_is_undef( struct ureg_dst dst ) { return dst.File == TGSI_FILE_NULL; diff --git a/src/gallium/auxiliary/translate/translate.h b/src/gallium/auxiliary/translate/translate.h index 7fe8ff8145f..d77561aa7ce 100644 --- a/src/gallium/auxiliary/translate/translate.h +++ b/src/gallium/auxiliary/translate/translate.h @@ -130,12 +130,12 @@ struct translate *translate_create( const struct translate_key *key ); boolean translate_is_output_format_supported(enum pipe_format format); -static INLINE int translate_keysize( const struct translate_key *key ) +static inline int translate_keysize( const struct translate_key *key ) { return 2 * sizeof(int) + key->nr_elements * sizeof(struct translate_element); } -static INLINE int translate_key_compare( const struct translate_key *a, +static inline int translate_key_compare( const struct translate_key *a, const struct translate_key *b ) { int keysize_a = translate_keysize(a); @@ -148,7 +148,7 @@ static INLINE int translate_key_compare( const struct translate_key *a, } -static INLINE void translate_key_sanitize( struct translate_key *a ) +static inline void translate_key_sanitize( struct translate_key *a ) { int keysize = translate_keysize(a); char *ptr = (char *)a; diff --git a/src/gallium/auxiliary/translate/translate_cache.c b/src/gallium/auxiliary/translate/translate_cache.c index bb8bdcb58c4..2bed02a454b 100644 --- a/src/gallium/auxiliary/translate/translate_cache.c +++ b/src/gallium/auxiliary/translate/translate_cache.c @@ -49,7 +49,7 @@ struct translate_cache * translate_cache_create( void ) } -static INLINE void delete_translates(struct translate_cache *cache) +static inline void delete_translates(struct translate_cache *cache) { struct cso_hash *hash = cache->hash; struct cso_hash_iter iter = cso_hash_first_node(hash); @@ -70,14 +70,14 @@ void translate_cache_destroy(struct translate_cache *cache) } -static INLINE unsigned translate_hash_key_size(struct translate_key *key) +static inline unsigned translate_hash_key_size(struct translate_key *key) { unsigned size = sizeof(struct translate_key) - sizeof(struct translate_element) * (TRANSLATE_MAX_ATTRIBS - key->nr_elements); return size; } -static INLINE unsigned create_key(struct translate_key *key) +static inline unsigned create_key(struct translate_key *key) { unsigned hash_key; unsigned size = translate_hash_key_size(key); diff --git a/src/gallium/auxiliary/util/u_bitmask.c b/src/gallium/auxiliary/util/u_bitmask.c index 23c93a3ebcb..b19be29a5a4 100644 --- a/src/gallium/auxiliary/util/u_bitmask.c +++ b/src/gallium/auxiliary/util/u_bitmask.c @@ -85,7 +85,7 @@ util_bitmask_create(void) /** * Resize the bitmask if necessary */ -static INLINE boolean +static inline boolean util_bitmask_resize(struct util_bitmask *bm, unsigned minimum_index) { @@ -131,7 +131,7 @@ util_bitmask_resize(struct util_bitmask *bm, /** * Lazily update the filled. */ -static INLINE void +static inline void util_bitmask_filled_set(struct util_bitmask *bm, unsigned index) { @@ -144,7 +144,7 @@ util_bitmask_filled_set(struct util_bitmask *bm, } } -static INLINE void +static inline void util_bitmask_filled_unset(struct util_bitmask *bm, unsigned index) { diff --git a/src/gallium/auxiliary/util/u_blend.h b/src/gallium/auxiliary/util/u_blend.h index 2485c34d418..4f969778972 100644 --- a/src/gallium/auxiliary/util/u_blend.h +++ b/src/gallium/auxiliary/util/u_blend.h @@ -9,7 +9,7 @@ * garbage that's there. Return a blend factor that will take that into * account. */ -static INLINE int +static inline int util_blend_dst_alpha_to_one(int factor) { switch (factor) { diff --git a/src/gallium/auxiliary/util/u_blit.c b/src/gallium/auxiliary/util/u_blit.c index e3f30557a03..9737c940936 100644 --- a/src/gallium/auxiliary/util/u_blit.c +++ b/src/gallium/auxiliary/util/u_blit.c @@ -158,7 +158,7 @@ util_destroy_blit(struct blit_state *ctx) /** * Helper function to set the fragment shaders. */ -static INLINE void +static inline void set_fragment_shader(struct blit_state *ctx, uint writemask, enum pipe_format format, enum pipe_texture_target pipe_tex) @@ -194,7 +194,7 @@ set_fragment_shader(struct blit_state *ctx, uint writemask, /** * Helper function to set the vertex shader. */ -static INLINE void +static inline void set_vertex_shader(struct blit_state *ctx) { /* vertex shader - still required to provide the linkage between @@ -546,8 +546,8 @@ util_blit_pixels_tex(struct blit_state *ctx, cso_save_rasterizer(ctx->cso); cso_save_sample_mask(ctx->cso); cso_save_min_samples(ctx->cso); - cso_save_samplers(ctx->cso, PIPE_SHADER_FRAGMENT); - cso_save_sampler_views(ctx->cso, PIPE_SHADER_FRAGMENT); + cso_save_fragment_samplers(ctx->cso); + cso_save_fragment_sampler_views(ctx->cso); cso_save_stream_outputs(ctx->cso); cso_save_viewport(ctx->cso); cso_save_framebuffer(ctx->cso); @@ -572,8 +572,10 @@ util_blit_pixels_tex(struct blit_state *ctx, ctx->sampler.normalized_coords = normalized; ctx->sampler.min_img_filter = filter; ctx->sampler.mag_img_filter = filter; - cso_single_sampler(ctx->cso, PIPE_SHADER_FRAGMENT, 0, &ctx->sampler); - cso_single_sampler_done(ctx->cso, PIPE_SHADER_FRAGMENT); + { + const struct pipe_sampler_state *samplers[] = {&ctx->sampler}; + cso_set_samplers(ctx->cso, PIPE_SHADER_FRAGMENT, 1, samplers); + } /* viewport */ ctx->viewport.scale[0] = 0.5f * dst->width; @@ -628,8 +630,8 @@ util_blit_pixels_tex(struct blit_state *ctx, cso_restore_rasterizer(ctx->cso); cso_restore_sample_mask(ctx->cso); cso_restore_min_samples(ctx->cso); - cso_restore_samplers(ctx->cso, PIPE_SHADER_FRAGMENT); - cso_restore_sampler_views(ctx->cso, PIPE_SHADER_FRAGMENT); + cso_restore_fragment_samplers(ctx->cso); + cso_restore_fragment_sampler_views(ctx->cso); cso_restore_viewport(ctx->cso); cso_restore_framebuffer(ctx->cso); cso_restore_fragment_shader(ctx->cso); diff --git a/src/gallium/auxiliary/util/u_blitter.c b/src/gallium/auxiliary/util/u_blitter.c index b5ef9a23966..85206eab1a7 100644 --- a/src/gallium/auxiliary/util/u_blitter.c +++ b/src/gallium/auxiliary/util/u_blitter.c @@ -938,7 +938,7 @@ static void *blitter_get_fs_texfetch_col(struct blitter_context_priv *ctx, } } -static INLINE +static inline void *blitter_get_fs_texfetch_depth(struct blitter_context_priv *ctx, enum pipe_texture_target target, unsigned nr_samples) @@ -976,7 +976,7 @@ void *blitter_get_fs_texfetch_depth(struct blitter_context_priv *ctx, } } -static INLINE +static inline void *blitter_get_fs_texfetch_depthstencil(struct blitter_context_priv *ctx, enum pipe_texture_target target, unsigned nr_samples) @@ -1014,7 +1014,7 @@ void *blitter_get_fs_texfetch_depthstencil(struct blitter_context_priv *ctx, } } -static INLINE +static inline void *blitter_get_fs_texfetch_stencil(struct blitter_context_priv *ctx, enum pipe_texture_target target, unsigned nr_samples) diff --git a/src/gallium/auxiliary/util/u_blitter.h b/src/gallium/auxiliary/util/u_blitter.h index 93b0e513bd0..0cd173d6284 100644 --- a/src/gallium/auxiliary/util/u_blitter.h +++ b/src/gallium/auxiliary/util/u_blitter.h @@ -143,7 +143,7 @@ void util_blitter_cache_all_shaders(struct blitter_context *blitter); /** * Return the pipe context associated with a blitter context. */ -static INLINE +static inline struct pipe_context *util_blitter_get_pipe(struct blitter_context *blitter) { return blitter->pipe; @@ -371,77 +371,77 @@ void util_blitter_custom_resolve_color(struct blitter_context *blitter, * * States not listed here are not affected by util_blitter. */ -static INLINE +static inline void util_blitter_save_blend(struct blitter_context *blitter, void *state) { blitter->saved_blend_state = state; } -static INLINE +static inline void util_blitter_save_depth_stencil_alpha(struct blitter_context *blitter, void *state) { blitter->saved_dsa_state = state; } -static INLINE +static inline void util_blitter_save_vertex_elements(struct blitter_context *blitter, void *state) { blitter->saved_velem_state = state; } -static INLINE +static inline void util_blitter_save_stencil_ref(struct blitter_context *blitter, const struct pipe_stencil_ref *state) { blitter->saved_stencil_ref = *state; } -static INLINE +static inline void util_blitter_save_rasterizer(struct blitter_context *blitter, void *state) { blitter->saved_rs_state = state; } -static INLINE +static inline void util_blitter_save_fragment_shader(struct blitter_context *blitter, void *fs) { blitter->saved_fs = fs; } -static INLINE +static inline void util_blitter_save_vertex_shader(struct blitter_context *blitter, void *vs) { blitter->saved_vs = vs; } -static INLINE +static inline void util_blitter_save_geometry_shader(struct blitter_context *blitter, void *gs) { blitter->saved_gs = gs; } -static INLINE void +static inline void util_blitter_save_tessctrl_shader(struct blitter_context *blitter, void *sh) { blitter->saved_tcs = sh; } -static INLINE void +static inline void util_blitter_save_tesseval_shader(struct blitter_context *blitter, void *sh) { blitter->saved_tes = sh; } -static INLINE +static inline void util_blitter_save_framebuffer(struct blitter_context *blitter, const struct pipe_framebuffer_state *state) { @@ -449,21 +449,21 @@ void util_blitter_save_framebuffer(struct blitter_context *blitter, util_copy_framebuffer_state(&blitter->saved_fb_state, state); } -static INLINE +static inline void util_blitter_save_viewport(struct blitter_context *blitter, struct pipe_viewport_state *state) { blitter->saved_viewport = *state; } -static INLINE +static inline void util_blitter_save_scissor(struct blitter_context *blitter, struct pipe_scissor_state *state) { blitter->saved_scissor = *state; } -static INLINE +static inline void util_blitter_save_fragment_sampler_states( struct blitter_context *blitter, unsigned num_sampler_states, @@ -476,7 +476,7 @@ void util_blitter_save_fragment_sampler_states( num_sampler_states * sizeof(void *)); } -static INLINE void +static inline void util_blitter_save_fragment_sampler_views(struct blitter_context *blitter, unsigned num_views, struct pipe_sampler_view **views) @@ -490,7 +490,7 @@ util_blitter_save_fragment_sampler_views(struct blitter_context *blitter, views[i]); } -static INLINE void +static inline void util_blitter_save_vertex_buffer_slot(struct blitter_context *blitter, struct pipe_vertex_buffer *vertex_buffers) { @@ -500,7 +500,7 @@ util_blitter_save_vertex_buffer_slot(struct blitter_context *blitter, sizeof(struct pipe_vertex_buffer)); } -static INLINE void +static inline void util_blitter_save_so_targets(struct blitter_context *blitter, unsigned num_targets, struct pipe_stream_output_target **targets) @@ -514,7 +514,7 @@ util_blitter_save_so_targets(struct blitter_context *blitter, targets[i]); } -static INLINE void +static inline void util_blitter_save_sample_mask(struct blitter_context *blitter, unsigned sample_mask) { @@ -522,7 +522,7 @@ util_blitter_save_sample_mask(struct blitter_context *blitter, blitter->saved_sample_mask = sample_mask; } -static INLINE void +static inline void util_blitter_save_render_condition(struct blitter_context *blitter, struct pipe_query *query, boolean condition, diff --git a/src/gallium/auxiliary/util/u_box.h b/src/gallium/auxiliary/util/u_box.h index 520a3d596cb..66cf989a830 100644 --- a/src/gallium/auxiliary/util/u_box.h +++ b/src/gallium/auxiliary/util/u_box.h @@ -4,7 +4,7 @@ #include "pipe/p_state.h" #include "util/u_math.h" -static INLINE +static inline void u_box_1d( unsigned x, unsigned w, struct pipe_box *box ) @@ -17,7 +17,7 @@ void u_box_1d( unsigned x, box->depth = 1; } -static INLINE +static inline void u_box_2d( unsigned x, unsigned y, unsigned w, @@ -32,7 +32,7 @@ void u_box_2d( unsigned x, box->depth = 1; } -static INLINE +static inline void u_box_origin_2d( unsigned w, unsigned h, struct pipe_box *box ) @@ -45,7 +45,7 @@ void u_box_origin_2d( unsigned w, box->depth = 1; } -static INLINE +static inline void u_box_2d_zslice( unsigned x, unsigned y, unsigned z, @@ -61,7 +61,7 @@ void u_box_2d_zslice( unsigned x, box->depth = 1; } -static INLINE +static inline void u_box_3d( unsigned x, unsigned y, unsigned z, @@ -86,7 +86,7 @@ void u_box_3d( unsigned x, * 3 if both width and height have been reduced. * Aliasing permitted. */ -static INLINE int +static inline int u_box_clip_2d(struct pipe_box *dst, const struct pipe_box *box, int w, int h) { @@ -129,14 +129,14 @@ u_box_clip_2d(struct pipe_box *dst, return res; } -static INLINE int64_t +static inline int64_t u_box_volume_3d(const struct pipe_box *box) { return (int64_t)box->width * box->height * box->depth; } /* Aliasing of @dst permitted. */ -static INLINE void +static inline void u_box_union_2d(struct pipe_box *dst, const struct pipe_box *a, const struct pipe_box *b) { @@ -148,7 +148,7 @@ u_box_union_2d(struct pipe_box *dst, } /* Aliasing of @dst permitted. */ -static INLINE void +static inline void u_box_union_3d(struct pipe_box *dst, const struct pipe_box *a, const struct pipe_box *b) { @@ -161,7 +161,7 @@ u_box_union_3d(struct pipe_box *dst, dst->depth = MAX2(a->z + a->depth, b->z + b->depth) - dst->z; } -static INLINE boolean +static inline boolean u_box_test_intersection_2d(const struct pipe_box *a, const struct pipe_box *b) { @@ -185,7 +185,7 @@ u_box_test_intersection_2d(const struct pipe_box *a, return TRUE; } -static INLINE void +static inline void u_box_minify_2d(struct pipe_box *dst, const struct pipe_box *src, unsigned l) { diff --git a/src/gallium/auxiliary/util/u_cache.c b/src/gallium/auxiliary/util/u_cache.c index 9395c66f2f8..da0856981eb 100644 --- a/src/gallium/auxiliary/util/u_cache.c +++ b/src/gallium/auxiliary/util/u_cache.c @@ -155,7 +155,7 @@ util_cache_entry_get(struct util_cache *cache, return NULL; } -static INLINE void +static inline void util_cache_entry_destroy(struct util_cache *cache, struct util_cache_entry *entry) { diff --git a/src/gallium/auxiliary/util/u_clear.h b/src/gallium/auxiliary/util/u_clear.h index af557be00bd..864d1302b4f 100644 --- a/src/gallium/auxiliary/util/u_clear.h +++ b/src/gallium/auxiliary/util/u_clear.h @@ -37,7 +37,7 @@ * Clear the given buffers to the specified values. * No masking, no scissor (clear entire buffer). */ -static INLINE void +static inline void util_clear(struct pipe_context *pipe, struct pipe_framebuffer_state *framebuffer, unsigned buffers, const union pipe_color_union *color, double depth, unsigned stencil) diff --git a/src/gallium/auxiliary/util/u_cpu_detect.c b/src/gallium/auxiliary/util/u_cpu_detect.c index 23ab46c54bc..d1f9e978682 100644 --- a/src/gallium/auxiliary/util/u_cpu_detect.c +++ b/src/gallium/auxiliary/util/u_cpu_detect.c @@ -179,7 +179,7 @@ static int has_cpuid(void) * @sa cpuid.h included in gcc-4.3 onwards. * @sa http://msdn.microsoft.com/en-us/library/hskdteyh.aspx */ -static INLINE void +static inline void cpuid(uint32_t ax, uint32_t *p) { #if (defined(PIPE_CC_GCC) || defined(PIPE_CC_SUNPRO)) && defined(PIPE_ARCH_X86) @@ -216,7 +216,7 @@ cpuid(uint32_t ax, uint32_t *p) * @sa cpuid.h included in gcc-4.4 onwards. * @sa http://msdn.microsoft.com/en-us/library/hskdteyh%28v=vs.90%29.aspx */ -static INLINE void +static inline void cpuid_count(uint32_t ax, uint32_t cx, uint32_t *p) { #if (defined(PIPE_CC_GCC) || defined(PIPE_CC_SUNPRO)) && defined(PIPE_ARCH_X86) @@ -250,7 +250,7 @@ cpuid_count(uint32_t ax, uint32_t cx, uint32_t *p) } -static INLINE uint64_t xgetbv(void) +static inline uint64_t xgetbv(void) { #if defined(PIPE_CC_GCC) uint32_t eax, edx; @@ -272,7 +272,7 @@ static INLINE uint64_t xgetbv(void) #if defined(PIPE_ARCH_X86) -PIPE_ALIGN_STACK static INLINE boolean sse2_has_daz(void) +PIPE_ALIGN_STACK static inline boolean sse2_has_daz(void) { struct { uint32_t pad1[7]; diff --git a/src/gallium/auxiliary/util/u_debug.c b/src/gallium/auxiliary/util/u_debug.c index 2d2d049b205..b4503deb8f6 100644 --- a/src/gallium/auxiliary/util/u_debug.c +++ b/src/gallium/auxiliary/util/u_debug.c @@ -256,12 +256,12 @@ static boolean str_has_option(const char *str, const char *name) return FALSE; } -unsigned long +uint64_t debug_get_flags_option(const char *name, const struct debug_named_value *flags, - unsigned long dfault) + uint64_t dfault) { - unsigned long result; + uint64_t result; const char *str; const struct debug_named_value *orig = flags; unsigned namealign = 0; @@ -276,7 +276,7 @@ debug_get_flags_option(const char *name, namealign = MAX2(namealign, strlen(flags->name)); for (flags = orig; flags->name; ++flags) _debug_printf("| %*s [0x%0*lx]%s%s\n", namealign, flags->name, - (int)sizeof(unsigned long)*CHAR_BIT/4, flags->value, + (int)sizeof(uint64_t)*CHAR_BIT/4, flags->value, flags->desc ? " " : "", flags->desc ? flags->desc : ""); } else { @@ -758,7 +758,8 @@ debug_print_bind_flags(const char *msg, unsigned usage) DEBUG_NAMED_VALUE(PIPE_BIND_CURSOR), DEBUG_NAMED_VALUE(PIPE_BIND_CUSTOM), DEBUG_NAMED_VALUE(PIPE_BIND_GLOBAL), - DEBUG_NAMED_VALUE(PIPE_BIND_SHADER_RESOURCE), + DEBUG_NAMED_VALUE(PIPE_BIND_SHADER_BUFFER), + DEBUG_NAMED_VALUE(PIPE_BIND_SHADER_IMAGE), DEBUG_NAMED_VALUE(PIPE_BIND_COMPUTE_RESOURCE), DEBUG_NAMED_VALUE(PIPE_BIND_COMMAND_ARGS_BUFFER), DEBUG_NAMED_VALUE(PIPE_BIND_SCANOUT), diff --git a/src/gallium/auxiliary/util/u_debug.h b/src/gallium/auxiliary/util/u_debug.h index 3b2255244a7..926063a1918 100644 --- a/src/gallium/auxiliary/util/u_debug.h +++ b/src/gallium/auxiliary/util/u_debug.h @@ -58,7 +58,7 @@ extern "C" { void _debug_vprintf(const char *format, va_list ap); -static INLINE void +static inline void _debug_printf(const char *format, ...) { va_list ap; @@ -78,10 +78,10 @@ _debug_printf(const char *format, ...) * that is guaranteed to be printed in all platforms) */ #if !defined(PIPE_OS_HAIKU) -static INLINE void +static inline void debug_printf(const char *format, ...) _util_printf_format(1,2); -static INLINE void +static inline void debug_printf(const char *format, ...) { #ifdef DEBUG @@ -269,7 +269,7 @@ void _debug_assert_fail(const char *expr, struct debug_named_value { const char *name; - unsigned long value; + uint64_t value; const char *desc; }; @@ -377,10 +377,10 @@ debug_get_bool_option(const char *name, boolean dfault); long debug_get_num_option(const char *name, long dfault); -unsigned long +uint64_t debug_get_flags_option(const char *name, const struct debug_named_value *flags, - unsigned long dfault); + uint64_t dfault); #define DEBUG_GET_ONCE_BOOL_OPTION(sufix, name, dfault) \ static boolean \ diff --git a/src/gallium/auxiliary/util/u_debug_describe.c b/src/gallium/auxiliary/util/u_debug_describe.c index df73ed83ef6..f428d22d205 100644 --- a/src/gallium/auxiliary/util/u_debug_describe.c +++ b/src/gallium/auxiliary/util/u_debug_describe.c @@ -81,6 +81,15 @@ debug_describe_sampler_view(char* buf, const struct pipe_sampler_view *ptr) } void +debug_describe_image_view(char* buf, const struct pipe_image_view *ptr) +{ + char res[128]; + debug_describe_resource(res, ptr->resource); + util_sprintf(buf, "pipe_image_view<%s,%s>", res, + util_format_short_name(ptr->format)); +} + +void debug_describe_so_target(char* buf, const struct pipe_stream_output_target *ptr) { diff --git a/src/gallium/auxiliary/util/u_debug_describe.h b/src/gallium/auxiliary/util/u_debug_describe.h index 4f7882b0b37..2172ecb4395 100644 --- a/src/gallium/auxiliary/util/u_debug_describe.h +++ b/src/gallium/auxiliary/util/u_debug_describe.h @@ -35,12 +35,14 @@ struct pipe_reference; struct pipe_resource; struct pipe_surface; struct pipe_sampler_view; +struct pipe_image_view; /* a 256-byte buffer is necessary and sufficient */ void debug_describe_reference(char* buf, const struct pipe_reference*ptr); void debug_describe_resource(char* buf, const struct pipe_resource *ptr); void debug_describe_surface(char* buf, const struct pipe_surface *ptr); void debug_describe_sampler_view(char* buf, const struct pipe_sampler_view *ptr); +void debug_describe_image_view(char* buf, const struct pipe_image_view *ptr); void debug_describe_so_target(char* buf, const struct pipe_stream_output_target *ptr); diff --git a/src/gallium/auxiliary/util/u_debug_memory.c b/src/gallium/auxiliary/util/u_debug_memory.c index 747837cd148..3e7ecfa79f3 100644 --- a/src/gallium/auxiliary/util/u_debug_memory.c +++ b/src/gallium/auxiliary/util/u_debug_memory.c @@ -92,7 +92,7 @@ pipe_static_mutex(list_mutex); static unsigned long last_no = 0; -static INLINE struct debug_memory_header * +static inline struct debug_memory_header * header_from_data(void *data) { if(data) @@ -101,7 +101,7 @@ header_from_data(void *data) return NULL; } -static INLINE void * +static inline void * data_from_header(struct debug_memory_header *hdr) { if(hdr) @@ -110,7 +110,7 @@ data_from_header(struct debug_memory_header *hdr) return NULL; } -static INLINE struct debug_memory_footer * +static inline struct debug_memory_footer * footer_from_header(struct debug_memory_header *hdr) { if(hdr) diff --git a/src/gallium/auxiliary/util/u_debug_refcnt.h b/src/gallium/auxiliary/util/u_debug_refcnt.h index c02fba27ddf..1f9218fec9a 100644 --- a/src/gallium/auxiliary/util/u_debug_refcnt.h +++ b/src/gallium/auxiliary/util/u_debug_refcnt.h @@ -42,7 +42,7 @@ extern int debug_refcnt_state; void debug_reference_slowpath(const struct pipe_reference* p, debug_reference_descriptor get_desc, int change); -static INLINE void debug_reference(const struct pipe_reference* p, debug_reference_descriptor get_desc, int change) +static inline void debug_reference(const struct pipe_reference* p, debug_reference_descriptor get_desc, int change) { if (debug_refcnt_state >= 0) debug_reference_slowpath(p, get_desc, change); @@ -50,7 +50,7 @@ static INLINE void debug_reference(const struct pipe_reference* p, debug_referen #else -static INLINE void debug_reference(const struct pipe_reference* p, debug_reference_descriptor get_desc, int change) +static inline void debug_reference(const struct pipe_reference* p, debug_reference_descriptor get_desc, int change) { } diff --git a/src/gallium/auxiliary/util/u_debug_symbol.c b/src/gallium/auxiliary/util/u_debug_symbol.c index 542493252ce..10efdd593e5 100644 --- a/src/gallium/auxiliary/util/u_debug_symbol.c +++ b/src/gallium/auxiliary/util/u_debug_symbol.c @@ -146,7 +146,7 @@ DBGHELP_DISPATCH(SymGetLineFromAddr64, #undef DBGHELP_DISPATCH -static INLINE boolean +static inline boolean debug_symbol_name_dbghelp(const void *addr, char* buf, unsigned size) { DWORD64 dwAddr = (DWORD64)(uintptr_t)addr; @@ -227,7 +227,7 @@ debug_symbol_name_dbghelp(const void *addr, char* buf, unsigned size) * * To fix this, post-process the output with tools/addr2line.sh */ -static INLINE boolean +static inline boolean debug_symbol_name_glibc(const void *addr, char* buf, unsigned size) { char** syms = backtrace_symbols((void**)&addr, 1); diff --git a/src/gallium/auxiliary/util/u_dirty_surfaces.h b/src/gallium/auxiliary/util/u_dirty_surfaces.h index d31f8b9170a..ccde8a8c115 100644 --- a/src/gallium/auxiliary/util/u_dirty_surfaces.h +++ b/src/gallium/auxiliary/util/u_dirty_surfaces.h @@ -47,13 +47,13 @@ struct util_dirty_surface struct list_head dirty_list; }; -static INLINE void +static inline void util_dirty_surfaces_init(struct util_dirty_surfaces *ds) { LIST_INITHEAD(&ds->dirty_list); } -static INLINE void +static inline void util_dirty_surfaces_use_for_sampling(struct pipe_context *pipe, struct util_dirty_surfaces *dss, util_dirty_surface_flush_t flush) { struct list_head *p, *next; @@ -66,7 +66,7 @@ util_dirty_surfaces_use_for_sampling(struct pipe_context *pipe, struct util_dirt } } -static INLINE void +static inline void util_dirty_surfaces_use_levels_for_sampling(struct pipe_context *pipe, struct util_dirty_surfaces *dss, unsigned first, unsigned last, util_dirty_surface_flush_t flush) { struct list_head *p, *next; @@ -82,7 +82,7 @@ util_dirty_surfaces_use_levels_for_sampling(struct pipe_context *pipe, struct ut } } -static INLINE void +static inline void util_dirty_surfaces_use_for_sampling_with(struct pipe_context *pipe, struct util_dirty_surfaces *dss, struct pipe_sampler_view *psv, struct pipe_sampler_state *pss, util_dirty_surface_flush_t flush) { if(!LIST_IS_EMPTY(&dss->dirty_list)) @@ -90,26 +90,26 @@ util_dirty_surfaces_use_for_sampling_with(struct pipe_context *pipe, struct util MIN2((unsigned)ceilf(pss->max_lod) + psv->u.tex.first_level, psv->u.tex.last_level), flush); } -static INLINE void +static inline void util_dirty_surface_init(struct util_dirty_surface *ds) { LIST_INITHEAD(&ds->dirty_list); } -static INLINE boolean +static inline boolean util_dirty_surface_is_dirty(struct util_dirty_surface *ds) { return !LIST_IS_EMPTY(&ds->dirty_list); } -static INLINE void +static inline void util_dirty_surface_set_dirty(struct util_dirty_surfaces *dss, struct util_dirty_surface *ds) { if(LIST_IS_EMPTY(&ds->dirty_list)) LIST_ADDTAIL(&ds->dirty_list, &dss->dirty_list); } -static INLINE void +static inline void util_dirty_surface_set_clean(struct util_dirty_surfaces *dss, struct util_dirty_surface *ds) { if(!LIST_IS_EMPTY(&ds->dirty_list)) diff --git a/src/gallium/auxiliary/util/u_draw.h b/src/gallium/auxiliary/util/u_draw.h index 9fc3e9924e1..5c0880f6ce4 100644 --- a/src/gallium/auxiliary/util/u_draw.h +++ b/src/gallium/auxiliary/util/u_draw.h @@ -39,7 +39,7 @@ extern "C" { #endif -static INLINE void +static inline void util_draw_init_info(struct pipe_draw_info *info) { memset(info, 0, sizeof(*info)); @@ -48,7 +48,7 @@ util_draw_init_info(struct pipe_draw_info *info) } -static INLINE void +static inline void util_draw_arrays(struct pipe_context *pipe, uint mode, uint start, uint count) { struct pipe_draw_info info; @@ -63,7 +63,7 @@ util_draw_arrays(struct pipe_context *pipe, uint mode, uint start, uint count) pipe->draw_vbo(pipe, &info); } -static INLINE void +static inline void util_draw_elements(struct pipe_context *pipe, int index_bias, uint mode, uint start, uint count) { @@ -79,7 +79,7 @@ util_draw_elements(struct pipe_context *pipe, int index_bias, pipe->draw_vbo(pipe, &info); } -static INLINE void +static inline void util_draw_arrays_instanced(struct pipe_context *pipe, uint mode, uint start, uint count, uint start_instance, @@ -99,7 +99,7 @@ util_draw_arrays_instanced(struct pipe_context *pipe, pipe->draw_vbo(pipe, &info); } -static INLINE void +static inline void util_draw_elements_instanced(struct pipe_context *pipe, int index_bias, uint mode, uint start, uint count, @@ -120,7 +120,7 @@ util_draw_elements_instanced(struct pipe_context *pipe, pipe->draw_vbo(pipe, &info); } -static INLINE void +static inline void util_draw_range_elements(struct pipe_context *pipe, int index_bias, uint min_index, diff --git a/src/gallium/auxiliary/util/u_dual_blend.h b/src/gallium/auxiliary/util/u_dual_blend.h index e31d43c18bd..9450800f715 100644 --- a/src/gallium/auxiliary/util/u_dual_blend.h +++ b/src/gallium/auxiliary/util/u_dual_blend.h @@ -3,7 +3,7 @@ #include "pipe/p_state.h" -static INLINE boolean util_blend_factor_is_dual_src(int factor) +static inline boolean util_blend_factor_is_dual_src(int factor) { return (factor == PIPE_BLENDFACTOR_SRC1_COLOR) || (factor == PIPE_BLENDFACTOR_SRC1_ALPHA) || @@ -11,7 +11,7 @@ static INLINE boolean util_blend_factor_is_dual_src(int factor) (factor == PIPE_BLENDFACTOR_INV_SRC1_ALPHA); } -static INLINE boolean util_blend_state_is_dual(const struct pipe_blend_state *blend, +static inline boolean util_blend_state_is_dual(const struct pipe_blend_state *blend, int index) { if (util_blend_factor_is_dual_src(blend->rt[index].rgb_src_factor) || diff --git a/src/gallium/auxiliary/util/u_dump.h b/src/gallium/auxiliary/util/u_dump.h index 58e7dfd8244..2598851152b 100644 --- a/src/gallium/auxiliary/util/u_dump.h +++ b/src/gallium/auxiliary/util/u_dump.h @@ -88,14 +88,16 @@ util_dump_tex_filter(unsigned value, boolean shortened); const char * util_dump_query_type(unsigned value, boolean shortened); +const char * +util_dump_prim_mode(unsigned value, boolean shortened); + /* * p_state.h, through a FILE */ void -util_dump_template(FILE *stream, - const struct pipe_resource *templat); +util_dump_resource(FILE *stream, const struct pipe_resource *state); void util_dump_rasterizer_state(FILE *stream, @@ -154,10 +156,23 @@ util_dump_surface(FILE *stream, const struct pipe_surface *state); void +util_dump_image_view(FILE *stream, const struct pipe_image_view *state); + +void +util_dump_sampler_view(FILE *stream, const struct pipe_sampler_view *state); + +void util_dump_transfer(FILE *stream, const struct pipe_transfer *state); void +util_dump_constant_buffer(FILE *stream, + const struct pipe_constant_buffer *state); + +void +util_dump_index_buffer(FILE *stream, const struct pipe_index_buffer *state); + +void util_dump_vertex_buffer(FILE *stream, const struct pipe_vertex_buffer *state); @@ -166,6 +181,10 @@ util_dump_vertex_element(FILE *stream, const struct pipe_vertex_element *state); void +util_dump_stream_output_target(FILE *stream, + const struct pipe_stream_output_target *state); + +void util_dump_draw_info(FILE *stream, const struct pipe_draw_info *state); void diff --git a/src/gallium/auxiliary/util/u_dump_defines.c b/src/gallium/auxiliary/util/u_dump_defines.c index 03fd15d0c44..3ddc9554b50 100644 --- a/src/gallium/auxiliary/util/u_dump_defines.c +++ b/src/gallium/auxiliary/util/u_dump_defines.c @@ -392,3 +392,44 @@ util_dump_query_type_short_names[] = { }; DEFINE_UTIL_DUMP_CONTINUOUS(query_type) + + +static const char * +util_dump_prim_mode_names[] = { + "PIPE_PRIM_POINTS", + "PIPE_PRIM_LINES", + "PIPE_PRIM_LINE_LOOP", + "PIPE_PRIM_LINE_STRIP", + "PIPE_PRIM_TRIANGLES", + "PIPE_PRIM_TRIANGLE_STRIP", + "PIPE_PRIM_TRIANGLE_FAN", + "PIPE_PRIM_QUADS", + "PIPE_PRIM_QUAD_STRIP", + "PIPE_PRIM_POLYGON", + "PIPE_PRIM_LINES_ADJACENCY", + "PIPE_PRIM_LINE_STRIP_ADJACENCY", + "PIPE_PRIM_TRIANGLES_ADJACENCY", + "PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY", + "PIPE_PRIM_PATCHES", +}; + +static const char * +util_dump_prim_mode_short_names[] = { + "points", + "lines", + "line_loop", + "line_strip", + "triangles", + "triangle_strip", + "triangle_fan", + "quads", + "quad_strip", + "polygon", + "lines_adjacency", + "line_strip_adjacency", + "triangles_adjacency", + "triangle_strip_adjacency", + "patches", +}; + +DEFINE_UTIL_DUMP_CONTINUOUS(prim_mode) diff --git a/src/gallium/auxiliary/util/u_dump_state.c b/src/gallium/auxiliary/util/u_dump_state.c index 7f620b50cf0..441d16236b5 100644 --- a/src/gallium/auxiliary/util/u_dump_state.c +++ b/src/gallium/auxiliary/util/u_dump_state.c @@ -39,7 +39,7 @@ * Dump primitives */ -static INLINE void +static inline void util_stream_writef(FILE *stream, const char *format, ...) { static char buf[1024]; @@ -247,6 +247,42 @@ util_dump_enum_func(FILE *stream, unsigned value) util_dump_enum(stream, util_dump_func(value, TRUE)); } +static void +util_dump_enum_prim_mode(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_prim_mode(value, TRUE)); +} + +static void +util_dump_enum_tex_target(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_tex_target(value, TRUE)); +} + +static void +util_dump_enum_tex_filter(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_tex_filter(value, TRUE)); +} + +static void +util_dump_enum_tex_mipfilter(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_tex_mipfilter(value, TRUE)); +} + +static void +util_dump_enum_tex_wrap(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_tex_wrap(value, TRUE)); +} + +static void +util_dump_enum_stencil_op(FILE *stream, unsigned value) +{ + util_dump_enum(stream, util_dump_stencil_op(value, TRUE)); +} + /* * Public functions @@ -254,38 +290,28 @@ util_dump_enum_func(FILE *stream, unsigned value) void -util_dump_template(FILE *stream, const struct pipe_resource *templat) +util_dump_resource(FILE *stream, const struct pipe_resource *state) { - if(!templat) { + if (!state) { util_dump_null(stream); return; } util_dump_struct_begin(stream, "pipe_resource"); - util_dump_member(stream, int, templat, target); - util_dump_member(stream, format, templat, format); - - util_dump_member_begin(stream, "width"); - util_dump_uint(stream, templat->width0); - util_dump_member_end(stream); - - util_dump_member_begin(stream, "height"); - util_dump_uint(stream, templat->height0); - util_dump_member_end(stream); - - util_dump_member_begin(stream, "depth"); - util_dump_uint(stream, templat->depth0); - util_dump_member_end(stream); + util_dump_member(stream, enum_tex_target, state, target); + util_dump_member(stream, format, state, format); - util_dump_member_begin(stream, "array_size"); - util_dump_uint(stream, templat->array_size); - util_dump_member_end(stream); + util_dump_member(stream, uint, state, width0); + util_dump_member(stream, uint, state, height0); + util_dump_member(stream, uint, state, depth0); + util_dump_member(stream, uint, state, array_size); - util_dump_member(stream, uint, templat, last_level); - util_dump_member(stream, uint, templat, usage); - util_dump_member(stream, uint, templat, bind); - util_dump_member(stream, uint, templat, flags); + util_dump_member(stream, uint, state, last_level); + util_dump_member(stream, uint, state, nr_samples); + util_dump_member(stream, uint, state, usage); + util_dump_member(stream, uint, state, bind); + util_dump_member(stream, uint, state, flags); util_dump_struct_end(stream); } @@ -319,6 +345,7 @@ util_dump_rasterizer_state(FILE *stream, const struct pipe_rasterizer_state *sta util_dump_member(stream, uint, state, sprite_coord_enable); util_dump_member(stream, bool, state, sprite_coord_mode); util_dump_member(stream, bool, state, point_quad_rasterization); + util_dump_member(stream, bool, state, point_tri_clip); util_dump_member(stream, bool, state, point_size_per_vertex); util_dump_member(stream, bool, state, multisample); util_dump_member(stream, bool, state, line_smooth); @@ -331,6 +358,7 @@ util_dump_rasterizer_state(FILE *stream, const struct pipe_rasterizer_state *sta util_dump_member(stream, bool, state, bottom_edge_rule); util_dump_member(stream, bool, state, rasterizer_discard); util_dump_member(stream, bool, state, depth_clip); + util_dump_member(stream, bool, state, clip_halfz); util_dump_member(stream, uint, state, clip_plane_enable); util_dump_member(stream, float, state, line_width); @@ -426,7 +454,6 @@ util_dump_clip_state(FILE *stream, const struct pipe_clip_state *state) void util_dump_shader_state(FILE *stream, const struct pipe_shader_state *state) { - char str[8192]; unsigned i; if(!state) { @@ -434,33 +461,35 @@ util_dump_shader_state(FILE *stream, const struct pipe_shader_state *state) return; } - tgsi_dump_str(state->tokens, 0, str, sizeof(str)); - util_dump_struct_begin(stream, "pipe_shader_state"); util_dump_member_begin(stream, "tokens"); - util_dump_string(stream, str); + fprintf(stream, "\"\n"); + tgsi_dump_to_file(state->tokens, 0, stream); + fprintf(stream, "\""); util_dump_member_end(stream); - util_dump_member_begin(stream, "stream_output"); - util_dump_struct_begin(stream, "pipe_stream_output_info"); - util_dump_member(stream, uint, &state->stream_output, num_outputs); - util_dump_array(stream, uint, state->stream_output.stride, - Elements(state->stream_output.stride)); - util_dump_array_begin(stream); - for(i = 0; i < state->stream_output.num_outputs; ++i) { - util_dump_elem_begin(stream); - util_dump_struct_begin(stream, ""); /* anonymous */ - util_dump_member(stream, uint, &state->stream_output.output[i], register_index); - util_dump_member(stream, uint, &state->stream_output.output[i], start_component); - util_dump_member(stream, uint, &state->stream_output.output[i], num_components); - util_dump_member(stream, uint, &state->stream_output.output[i], output_buffer); + if (state->stream_output.num_outputs) { + util_dump_member_begin(stream, "stream_output"); + util_dump_struct_begin(stream, "pipe_stream_output_info"); + util_dump_member(stream, uint, &state->stream_output, num_outputs); + util_dump_array(stream, uint, state->stream_output.stride, + Elements(state->stream_output.stride)); + util_dump_array_begin(stream); + for(i = 0; i < state->stream_output.num_outputs; ++i) { + util_dump_elem_begin(stream); + util_dump_struct_begin(stream, ""); /* anonymous */ + util_dump_member(stream, uint, &state->stream_output.output[i], register_index); + util_dump_member(stream, uint, &state->stream_output.output[i], start_component); + util_dump_member(stream, uint, &state->stream_output.output[i], num_components); + util_dump_member(stream, uint, &state->stream_output.output[i], output_buffer); + util_dump_struct_end(stream); + util_dump_elem_end(stream); + } + util_dump_array_end(stream); util_dump_struct_end(stream); - util_dump_elem_end(stream); + util_dump_member_end(stream); } - util_dump_array_end(stream); - util_dump_struct_end(stream); - util_dump_member_end(stream); util_dump_struct_end(stream); } @@ -496,9 +525,12 @@ util_dump_depth_stencil_alpha_state(FILE *stream, const struct pipe_depth_stenci util_dump_member(stream, bool, &state->stencil[i], enabled); if (state->stencil[i].enabled) { util_dump_member(stream, enum_func, &state->stencil[i], func); - util_dump_member(stream, uint, &state->stencil[i], fail_op); - util_dump_member(stream, uint, &state->stencil[i], zpass_op); - util_dump_member(stream, uint, &state->stencil[i], zfail_op); + util_dump_member(stream, enum_stencil_op, + &state->stencil[i], fail_op); + util_dump_member(stream, enum_stencil_op, + &state->stencil[i], zpass_op); + util_dump_member(stream, enum_stencil_op, + &state->stencil[i], zfail_op); util_dump_member(stream, uint, &state->stencil[i], valuemask); util_dump_member(stream, uint, &state->stencil[i], writemask); } @@ -555,6 +587,8 @@ util_dump_blend_state(FILE *stream, const struct pipe_blend_state *state) util_dump_struct_begin(stream, "pipe_blend_state"); util_dump_member(stream, bool, state, dither); + util_dump_member(stream, bool, state, alpha_to_coverage); + util_dump_member(stream, bool, state, alpha_to_one); util_dump_member(stream, bool, state, logicop_enable); if (state->logicop_enable) { @@ -629,16 +663,17 @@ util_dump_sampler_state(FILE *stream, const struct pipe_sampler_state *state) util_dump_struct_begin(stream, "pipe_sampler_state"); - util_dump_member(stream, uint, state, wrap_s); - util_dump_member(stream, uint, state, wrap_t); - util_dump_member(stream, uint, state, wrap_r); - util_dump_member(stream, uint, state, min_img_filter); - util_dump_member(stream, uint, state, min_mip_filter); - util_dump_member(stream, uint, state, mag_img_filter); + util_dump_member(stream, enum_tex_wrap, state, wrap_s); + util_dump_member(stream, enum_tex_wrap, state, wrap_t); + util_dump_member(stream, enum_tex_wrap, state, wrap_r); + util_dump_member(stream, enum_tex_filter, state, min_img_filter); + util_dump_member(stream, enum_tex_mipfilter, state, min_mip_filter); + util_dump_member(stream, enum_tex_filter, state, mag_img_filter); util_dump_member(stream, uint, state, compare_mode); util_dump_member(stream, enum_func, state, compare_func); util_dump_member(stream, bool, state, normalized_coords); util_dump_member(stream, uint, state, max_anisotropy); + util_dump_member(stream, bool, state, seamless_cube_map); util_dump_member(stream, float, state, lod_bias); util_dump_member(stream, float, state, min_lod); util_dump_member(stream, float, state, max_lod); @@ -672,6 +707,67 @@ util_dump_surface(FILE *stream, const struct pipe_surface *state) void +util_dump_image_view(FILE *stream, const struct pipe_image_view *state) +{ + if (!state) { + util_dump_null(stream); + return; + } + + util_dump_struct_begin(stream, "pipe_image_view"); + + util_dump_member(stream, ptr, state, resource); + util_dump_member(stream, format, state, format); + + if (state->resource->target == PIPE_BUFFER) { + util_dump_member(stream, uint, state, u.buf.first_element); + util_dump_member(stream, uint, state, u.buf.last_element); + } + else { + util_dump_member(stream, uint, state, u.tex.first_layer); + util_dump_member(stream, uint, state, u.tex.last_layer); + util_dump_member(stream, uint, state, u.tex.level); + } + + util_dump_struct_end(stream); +} + + +void +util_dump_sampler_view(FILE *stream, const struct pipe_sampler_view *state) +{ + if (!state) { + util_dump_null(stream); + return; + } + + util_dump_struct_begin(stream, "pipe_sampler_view"); + + util_dump_member(stream, enum_tex_target, state, target); + util_dump_member(stream, format, state, format); + util_dump_member(stream, ptr, state, texture); + + if (state->target == PIPE_BUFFER) { + util_dump_member(stream, uint, state, u.buf.first_element); + util_dump_member(stream, uint, state, u.buf.last_element); + } + else { + util_dump_member(stream, uint, state, u.tex.first_layer); + util_dump_member(stream, uint, state, u.tex.last_layer); + util_dump_member(stream, uint, state, u.tex.last_level); + util_dump_member(stream, uint, state, u.tex.last_level); + } + + util_dump_member(stream, uint, state, swizzle_r); + util_dump_member(stream, uint, state, swizzle_g); + util_dump_member(stream, uint, state, swizzle_b); + util_dump_member(stream, uint, state, swizzle_a); + + util_dump_struct_end(stream); +} + + +void util_dump_transfer(FILE *stream, const struct pipe_transfer *state) { if(!state) { @@ -695,6 +791,45 @@ util_dump_transfer(FILE *stream, const struct pipe_transfer *state) void +util_dump_constant_buffer(FILE *stream, + const struct pipe_constant_buffer *state) +{ + if (!state) { + util_dump_null(stream); + return; + } + + util_dump_struct_begin(stream, "pipe_constant_buffer"); + + util_dump_member(stream, ptr, state, buffer); + util_dump_member(stream, uint, state, buffer_offset); + util_dump_member(stream, uint, state, buffer_size); + util_dump_member(stream, ptr, state, user_buffer); + + util_dump_struct_end(stream); +} + + +void +util_dump_index_buffer(FILE *stream, const struct pipe_index_buffer *state) +{ + if (!state) { + util_dump_null(stream); + return; + } + + util_dump_struct_begin(stream, "pipe_index_buffer"); + + util_dump_member(stream, uint, state, index_size); + util_dump_member(stream, uint, state, offset); + util_dump_member(stream, ptr, state, buffer); + util_dump_member(stream, ptr, state, user_buffer); + + util_dump_struct_end(stream); +} + + +void util_dump_vertex_buffer(FILE *stream, const struct pipe_vertex_buffer *state) { if(!state) { @@ -707,6 +842,7 @@ util_dump_vertex_buffer(FILE *stream, const struct pipe_vertex_buffer *state) util_dump_member(stream, uint, state, stride); util_dump_member(stream, uint, state, buffer_offset); util_dump_member(stream, ptr, state, buffer); + util_dump_member(stream, ptr, state, user_buffer); util_dump_struct_end(stream); } @@ -732,6 +868,25 @@ util_dump_vertex_element(FILE *stream, const struct pipe_vertex_element *state) void +util_dump_stream_output_target(FILE *stream, + const struct pipe_stream_output_target *state) +{ + if (!state) { + util_dump_null(stream); + return; + } + + util_dump_struct_begin(stream, "pipe_stream_output_target"); + + util_dump_member(stream, ptr, state, buffer); + util_dump_member(stream, uint, state, buffer_offset); + util_dump_member(stream, uint, state, buffer_size); + + util_dump_struct_end(stream); +} + + +void util_dump_draw_info(FILE *stream, const struct pipe_draw_info *state) { if(!state) { @@ -743,7 +898,7 @@ util_dump_draw_info(FILE *stream, const struct pipe_draw_info *state) util_dump_member(stream, bool, state, indexed); - util_dump_member(stream, uint, state, mode); + util_dump_member(stream, enum_prim_mode, state, mode); util_dump_member(stream, uint, state, start); util_dump_member(stream, uint, state, count); @@ -830,12 +985,14 @@ void util_dump_blit_info(FILE *stream, const struct pipe_blit_info *info) util_dump_member_begin(stream, "mask"); util_dump_string(stream, mask); util_dump_member_end(stream); - util_dump_member(stream, uint, info, filter); + util_dump_member(stream, enum_tex_filter, info, filter); util_dump_member(stream, bool, info, scissor_enable); util_dump_member_begin(stream, "scissor"); util_dump_scissor_state(stream, &info->scissor); util_dump_member_end(stream); + util_dump_member(stream, bool, info, render_condition_enable); + util_dump_struct_end(stream); } diff --git a/src/gallium/auxiliary/util/u_dynarray.h b/src/gallium/auxiliary/util/u_dynarray.h index 980cadf22d1..7b7a093d824 100644 --- a/src/gallium/auxiliary/util/u_dynarray.h +++ b/src/gallium/auxiliary/util/u_dynarray.h @@ -43,13 +43,13 @@ struct util_dynarray unsigned capacity; }; -static INLINE void +static inline void util_dynarray_init(struct util_dynarray *buf) { memset(buf, 0, sizeof(*buf)); } -static INLINE void +static inline void util_dynarray_fini(struct util_dynarray *buf) { if(buf->data) @@ -60,7 +60,7 @@ util_dynarray_fini(struct util_dynarray *buf) } /* use util_dynarray_trim to reduce the allocated storage */ -static INLINE void * +static inline void * util_dynarray_resize(struct util_dynarray *buf, unsigned newsize) { char *p; @@ -78,13 +78,13 @@ util_dynarray_resize(struct util_dynarray *buf, unsigned newsize) return p; } -static INLINE void * +static inline void * util_dynarray_grow(struct util_dynarray *buf, int diff) { return util_dynarray_resize(buf, buf->size + diff); } -static INLINE void +static inline void util_dynarray_trim(struct util_dynarray *buf) { if (buf->size != buf->capacity) { diff --git a/src/gallium/auxiliary/util/u_fifo.h b/src/gallium/auxiliary/util/u_fifo.h index 9e007de1ada..a7aad6179d9 100644 --- a/src/gallium/auxiliary/util/u_fifo.h +++ b/src/gallium/auxiliary/util/u_fifo.h @@ -36,7 +36,7 @@ struct util_fifo size_t size; }; -static INLINE struct util_fifo * +static inline struct util_fifo * u_fifo_create(size_t size) { struct util_fifo *fifo; @@ -50,7 +50,7 @@ u_fifo_create(size_t size) return fifo; } -static INLINE boolean +static inline boolean u_fifo_add(struct util_fifo *fifo, void *ptr) { void **array = (void**)&fifo[1]; @@ -67,7 +67,7 @@ u_fifo_add(struct util_fifo *fifo, void *ptr) return TRUE; } -static INLINE boolean +static inline boolean u_fifo_pop(struct util_fifo *fifo, void **ptr) { void **array = (void**)&fifo[1]; @@ -85,7 +85,7 @@ u_fifo_pop(struct util_fifo *fifo, void **ptr) return TRUE; } -static INLINE void +static inline void u_fifo_destroy(struct util_fifo *fifo) { FREE(fifo); diff --git a/src/gallium/auxiliary/util/u_format.h b/src/gallium/auxiliary/util/u_format.h index 621574c9673..42b39ff04fd 100644 --- a/src/gallium/auxiliary/util/u_format.h +++ b/src/gallium/auxiliary/util/u_format.h @@ -425,7 +425,7 @@ util_format_description(enum pipe_format format); * Format query functions. */ -static INLINE const char * +static inline const char * util_format_name(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -438,7 +438,7 @@ util_format_name(enum pipe_format format) return desc->name; } -static INLINE const char * +static inline const char * util_format_short_name(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -454,7 +454,7 @@ util_format_short_name(enum pipe_format format) /** * Whether this format is plain, see UTIL_FORMAT_LAYOUT_PLAIN for more info. */ -static INLINE boolean +static inline boolean util_format_is_plain(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -466,7 +466,7 @@ util_format_is_plain(enum pipe_format format) return desc->layout == UTIL_FORMAT_LAYOUT_PLAIN ? TRUE : FALSE; } -static INLINE boolean +static inline boolean util_format_is_compressed(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -488,7 +488,7 @@ util_format_is_compressed(enum pipe_format format) } } -static INLINE boolean +static inline boolean util_format_is_s3tc(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -501,28 +501,28 @@ util_format_is_s3tc(enum pipe_format format) return desc->layout == UTIL_FORMAT_LAYOUT_S3TC ? TRUE : FALSE; } -static INLINE boolean +static inline boolean util_format_is_srgb(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); return desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB; } -static INLINE boolean +static inline boolean util_format_has_depth(const struct util_format_description *desc) { return desc->colorspace == UTIL_FORMAT_COLORSPACE_ZS && desc->swizzle[0] != UTIL_FORMAT_SWIZZLE_NONE; } -static INLINE boolean +static inline boolean util_format_has_stencil(const struct util_format_description *desc) { return desc->colorspace == UTIL_FORMAT_COLORSPACE_ZS && desc->swizzle[1] != UTIL_FORMAT_SWIZZLE_NONE; } -static INLINE boolean +static inline boolean util_format_is_depth_or_stencil(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -536,7 +536,7 @@ util_format_is_depth_or_stencil(enum pipe_format format) util_format_has_stencil(desc); } -static INLINE boolean +static inline boolean util_format_is_depth_and_stencil(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -554,7 +554,7 @@ util_format_is_depth_and_stencil(enum pipe_format format) /** * Calculates the depth format type based upon the incoming format description. */ -static INLINE unsigned +static inline unsigned util_get_depth_format_type(const struct util_format_description *desc) { unsigned depth_channel = desc->swizzle[0]; @@ -581,7 +581,7 @@ util_get_depth_format_mrd(const struct util_format_description *desc); * Return whether this is an RGBA, Z, S, or combined ZS format. * Useful for initializing pipe_blit_info::mask. */ -static INLINE unsigned +static inline unsigned util_format_get_mask(enum pipe_format format) { const struct util_format_description *desc = @@ -611,7 +611,7 @@ util_format_get_mask(enum pipe_format format) * * That is, the channels whose values are preserved. */ -static INLINE unsigned +static inline unsigned util_format_colormask(const struct util_format_description *desc) { unsigned colormask; @@ -643,7 +643,7 @@ util_format_colormask(const struct util_format_description *desc) * @param desc a format description to check colormask with * @param colormask a bit mask for channels, matches format of PIPE_MASK_RGBA */ -static INLINE boolean +static inline boolean util_format_colormask_full(const struct util_format_description *desc, unsigned colormask) { return (~colormask & util_format_colormask(desc)) == 0; @@ -709,7 +709,7 @@ util_format_is_supported(enum pipe_format format, unsigned bind); * * PIPE_FORMAT_?8?8?8?8_UNORM */ -static INLINE boolean +static inline boolean util_format_is_rgba8_variant(const struct util_format_description *desc) { unsigned chan; @@ -737,7 +737,7 @@ util_format_is_rgba8_variant(const struct util_format_description *desc) /** * Return total bits needed for the pixel format per block. */ -static INLINE uint +static inline uint util_format_get_blocksizebits(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -753,7 +753,7 @@ util_format_get_blocksizebits(enum pipe_format format) /** * Return bytes per block (not pixel) for the given format. */ -static INLINE uint +static inline uint util_format_get_blocksize(enum pipe_format format) { uint bits = util_format_get_blocksizebits(format); @@ -768,7 +768,7 @@ util_format_get_blocksize(enum pipe_format format) return bytes; } -static INLINE uint +static inline uint util_format_get_blockwidth(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -781,7 +781,7 @@ util_format_get_blockwidth(enum pipe_format format) return desc->block.width; } -static INLINE uint +static inline uint util_format_get_blockheight(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -794,7 +794,7 @@ util_format_get_blockheight(enum pipe_format format) return desc->block.height; } -static INLINE unsigned +static inline unsigned util_format_get_nblocksx(enum pipe_format format, unsigned x) { @@ -802,7 +802,7 @@ util_format_get_nblocksx(enum pipe_format format, return (x + blockwidth - 1) / blockwidth; } -static INLINE unsigned +static inline unsigned util_format_get_nblocksy(enum pipe_format format, unsigned y) { @@ -810,7 +810,7 @@ util_format_get_nblocksy(enum pipe_format format, return (y + blockheight - 1) / blockheight; } -static INLINE unsigned +static inline unsigned util_format_get_nblocks(enum pipe_format format, unsigned width, unsigned height) @@ -818,14 +818,14 @@ util_format_get_nblocks(enum pipe_format format, return util_format_get_nblocksx(format, width) * util_format_get_nblocksy(format, height); } -static INLINE size_t +static inline size_t util_format_get_stride(enum pipe_format format, unsigned width) { return util_format_get_nblocksx(format, width) * util_format_get_blocksize(format); } -static INLINE size_t +static inline size_t util_format_get_2d_size(enum pipe_format format, size_t stride, unsigned height) @@ -833,7 +833,7 @@ util_format_get_2d_size(enum pipe_format format, return util_format_get_nblocksy(format, height) * stride; } -static INLINE uint +static inline uint util_format_get_component_bits(enum pipe_format format, enum util_format_colorspace colorspace, uint component) @@ -880,7 +880,7 @@ util_format_get_component_bits(enum pipe_format format, * Given a linear RGB colorspace format, return the corresponding SRGB * format, or PIPE_FORMAT_NONE if none. */ -static INLINE enum pipe_format +static inline enum pipe_format util_format_srgb(enum pipe_format format) { if (util_format_is_srgb(format)) @@ -930,7 +930,7 @@ util_format_srgb(enum pipe_format format) * Given an sRGB format, return the corresponding linear colorspace format. * For non sRGB formats, return the format unchanged. */ -static INLINE enum pipe_format +static inline enum pipe_format util_format_linear(enum pipe_format format) { switch (format) { @@ -977,7 +977,7 @@ util_format_linear(enum pipe_format format) * Given a depth-stencil format, return the corresponding stencil-only format. * For stencil-only formats, return the format unchanged. */ -static INLINE enum pipe_format +static inline enum pipe_format util_format_stencil_only(enum pipe_format format) { switch (format) { @@ -1006,7 +1006,7 @@ util_format_stencil_only(enum pipe_format format) * Converts PIPE_FORMAT_*I* to PIPE_FORMAT_*R*. * This is identity for non-intensity formats. */ -static INLINE enum pipe_format +static inline enum pipe_format util_format_intensity_to_red(enum pipe_format format) { switch (format) { @@ -1044,7 +1044,7 @@ util_format_intensity_to_red(enum pipe_format format) * Converts PIPE_FORMAT_*L* to PIPE_FORMAT_*R*. * This is identity for non-luminance formats. */ -static INLINE enum pipe_format +static inline enum pipe_format util_format_luminance_to_red(enum pipe_format format) { switch (format) { @@ -1122,7 +1122,7 @@ util_format_luminance_to_red(enum pipe_format format) * Return the number of components stored. * Formats with block size != 1x1 will always have 1 component (the block). */ -static INLINE unsigned +static inline unsigned util_format_get_nr_components(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); @@ -1133,7 +1133,7 @@ util_format_get_nr_components(enum pipe_format format) * Return the index of the first non-void channel * -1 if no non-void channels */ -static INLINE int +static inline int util_format_get_first_non_void_channel(enum pipe_format format) { const struct util_format_description *desc = util_format_description(format); diff --git a/src/gallium/auxiliary/util/u_format_pack.py b/src/gallium/auxiliary/util/u_format_pack.py index d5138cc0577..fb42de723c4 100644 --- a/src/gallium/auxiliary/util/u_format_pack.py +++ b/src/gallium/auxiliary/util/u_format_pack.py @@ -616,7 +616,7 @@ def generate_format_unpack(format, dst_channel, dst_native_type, dst_suffix): name = format.short_name() - print 'static INLINE void' + print 'static inline void' print 'util_format_%s_unpack_%s(%s *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)' % (name, dst_suffix, dst_native_type) print '{' @@ -645,7 +645,7 @@ def generate_format_pack(format, src_channel, src_native_type, src_suffix): name = format.short_name() - print 'static INLINE void' + print 'static inline void' print 'util_format_%s_pack_%s(uint8_t *dst_row, unsigned dst_stride, const %s *src_row, unsigned src_stride, unsigned width, unsigned height)' % (name, src_suffix, src_native_type) print '{' @@ -674,7 +674,7 @@ def generate_format_fetch(format, dst_channel, dst_native_type, dst_suffix): name = format.short_name() - print 'static INLINE void' + print 'static inline void' print 'util_format_%s_fetch_%s(%s *dst, const uint8_t *src, unsigned i, unsigned j)' % (name, dst_suffix, dst_native_type) print '{' diff --git a/src/gallium/auxiliary/util/u_format_r11g11b10f.h b/src/gallium/auxiliary/util/u_format_r11g11b10f.h index 57516c39c6e..218822b16e6 100644 --- a/src/gallium/auxiliary/util/u_format_r11g11b10f.h +++ b/src/gallium/auxiliary/util/u_format_r11g11b10f.h @@ -45,7 +45,7 @@ #define F32_INFINITY 0x7f800000 -static INLINE unsigned f32_to_uf11(float val) +static inline unsigned f32_to_uf11(float val) { union { float f; @@ -94,7 +94,7 @@ static INLINE unsigned f32_to_uf11(float val) return uf11; } -static INLINE float uf11_to_f32(uint16_t val) +static inline float uf11_to_f32(uint16_t val) { union { float f; @@ -131,7 +131,7 @@ static INLINE float uf11_to_f32(uint16_t val) return f32.f; } -static INLINE unsigned f32_to_uf10(float val) +static inline unsigned f32_to_uf10(float val) { union { float f; @@ -180,7 +180,7 @@ static INLINE unsigned f32_to_uf10(float val) return uf10; } -static INLINE float uf10_to_f32(uint16_t val) +static inline float uf10_to_f32(uint16_t val) { union { float f; @@ -217,14 +217,14 @@ static INLINE float uf10_to_f32(uint16_t val) return f32.f; } -static INLINE unsigned float3_to_r11g11b10f(const float rgb[3]) +static inline unsigned float3_to_r11g11b10f(const float rgb[3]) { return ( f32_to_uf11(rgb[0]) & 0x7ff) | ((f32_to_uf11(rgb[1]) & 0x7ff) << 11) | ((f32_to_uf10(rgb[2]) & 0x3ff) << 22); } -static INLINE void r11g11b10f_to_float3(unsigned rgb, float retval[3]) +static inline void r11g11b10f_to_float3(unsigned rgb, float retval[3]) { retval[0] = uf11_to_f32( rgb & 0x7ff); retval[1] = uf11_to_f32((rgb >> 11) & 0x7ff); diff --git a/src/gallium/auxiliary/util/u_format_rgb9e5.h b/src/gallium/auxiliary/util/u_format_rgb9e5.h index c2a3f6f3e9d..59fc291e917 100644 --- a/src/gallium/auxiliary/util/u_format_rgb9e5.h +++ b/src/gallium/auxiliary/util/u_format_rgb9e5.h @@ -26,9 +26,10 @@ #ifndef RGB9E5_H #define RGB9E5_H -#include <math.h> #include <assert.h> +#include "c99_math.h" + #define RGB9E5_EXPONENT_BITS 5 #define RGB9E5_MANTISSA_BITS 9 #define RGB9E5_EXP_BIAS 15 @@ -73,9 +74,9 @@ typedef union { } field; } rgb9e5; -static INLINE float rgb9e5_ClampRange(float x) +static inline float rgb9e5_ClampRange(float x) { - if (x > 0.0) { + if (x > 0.0f) { if (x >= MAX_RGB9E5) { return MAX_RGB9E5; } else { @@ -90,7 +91,7 @@ static INLINE float rgb9e5_ClampRange(float x) /* Ok, FloorLog2 is not correct for the denorm and zero values, but we are going to do a max of this value with the minimum rgb9e5 exponent that will hide these problem cases. */ -static INLINE int rgb9e5_FloorLog2(float x) +static inline int rgb9e5_FloorLog2(float x) { float754 f; @@ -98,7 +99,7 @@ static INLINE int rgb9e5_FloorLog2(float x) return (f.field.biasedexponent - 127); } -static INLINE unsigned float3_to_rgb9e5(const float rgb[3]) +static inline unsigned float3_to_rgb9e5(const float rgb[3]) { rgb9e5 retval; float maxrgb; @@ -115,8 +116,8 @@ static INLINE unsigned float3_to_rgb9e5(const float rgb[3]) exp_shared = MAX2(-RGB9E5_EXP_BIAS-1, rgb9e5_FloorLog2(maxrgb)) + 1 + RGB9E5_EXP_BIAS; assert(exp_shared <= RGB9E5_MAX_VALID_BIASED_EXP); assert(exp_shared >= 0); - /* This pow function could be replaced by a table. */ - denom = pow(2, exp_shared - RGB9E5_EXP_BIAS - RGB9E5_MANTISSA_BITS); + /* This exp2 function could be replaced by a table. */ + denom = exp2(exp_shared - RGB9E5_EXP_BIAS - RGB9E5_MANTISSA_BITS); maxm = (int) floor(maxrgb / denom + 0.5); if (maxm == MAX_RGB9E5_MANTISSA+1) { @@ -146,7 +147,7 @@ static INLINE unsigned float3_to_rgb9e5(const float rgb[3]) return retval.raw; } -static INLINE void rgb9e5_to_float3(unsigned rgb, float retval[3]) +static inline void rgb9e5_to_float3(unsigned rgb, float retval[3]) { rgb9e5 v; int exponent; @@ -154,7 +155,7 @@ static INLINE void rgb9e5_to_float3(unsigned rgb, float retval[3]) v.raw = rgb; exponent = v.field.biasedexponent - RGB9E5_EXP_BIAS - RGB9E5_MANTISSA_BITS; - scale = (float) pow(2, exponent); + scale = exp2f(exponent); retval[0] = v.field.r * scale; retval[1] = v.field.g * scale; diff --git a/src/gallium/auxiliary/util/u_format_s3tc.c b/src/gallium/auxiliary/util/u_format_s3tc.c index 7e05989e6a1..cd3e165d3f0 100644 --- a/src/gallium/auxiliary/util/u_format_s3tc.c +++ b/src/gallium/auxiliary/util/u_format_s3tc.c @@ -235,7 +235,7 @@ util_format_dxt5_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned * Block decompression. */ -static INLINE void +static inline void util_format_dxtn_rgb_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height, @@ -312,7 +312,7 @@ util_format_dxt5_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, 16, FALSE); } -static INLINE void +static inline void util_format_dxtn_rgb_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height, @@ -400,7 +400,7 @@ util_format_dxt5_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, * Block compression. */ -static INLINE void +static inline void util_format_dxtn_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src, unsigned src_stride, unsigned width, unsigned height, @@ -478,7 +478,7 @@ util_format_dxt5_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, 16, FALSE); } -static INLINE void +static inline void util_format_dxtn_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src, unsigned src_stride, unsigned width, unsigned height, diff --git a/src/gallium/auxiliary/util/u_format_yuv.h b/src/gallium/auxiliary/util/u_format_yuv.h index 4ec39812e47..41524d63f3a 100644 --- a/src/gallium/auxiliary/util/u_format_yuv.h +++ b/src/gallium/auxiliary/util/u_format_yuv.h @@ -54,7 +54,7 @@ * precision in the coefficients. */ -static INLINE void +static inline void util_format_rgb_float_to_yuv(float r, float g, float b, uint8_t *y, uint8_t *u, uint8_t *v) { @@ -74,7 +74,7 @@ util_format_rgb_float_to_yuv(float r, float g, float b, } -static INLINE void +static inline void util_format_yuv_to_rgb_float(uint8_t y, uint8_t u, uint8_t v, float *r, float *g, float *b) { @@ -92,7 +92,7 @@ util_format_yuv_to_rgb_float(uint8_t y, uint8_t u, uint8_t v, } -static INLINE void +static inline void util_format_rgb_8unorm_to_yuv(uint8_t r, uint8_t g, uint8_t b, uint8_t *y, uint8_t *u, uint8_t *v) { @@ -102,7 +102,7 @@ util_format_rgb_8unorm_to_yuv(uint8_t r, uint8_t g, uint8_t b, } -static INLINE void +static inline void util_format_yuv_to_rgb_8unorm(uint8_t y, uint8_t u, uint8_t v, uint8_t *r, uint8_t *g, uint8_t *b) { diff --git a/src/gallium/auxiliary/util/u_format_zs.c b/src/gallium/auxiliary/util/u_format_zs.c index f1ed32f1d5c..69f2f2971f7 100644 --- a/src/gallium/auxiliary/util/u_format_zs.c +++ b/src/gallium/auxiliary/util/u_format_zs.c @@ -35,28 +35,28 @@ * z32_unorm conversion functions */ -static INLINE uint16_t +static inline uint16_t z32_unorm_to_z16_unorm(uint32_t z) { /* z * 0xffff / 0xffffffff */ return z >> 16; } -static INLINE uint32_t +static inline uint32_t z16_unorm_to_z32_unorm(uint16_t z) { /* z * 0xffffffff / 0xffff */ return (z << 16) | z; } -static INLINE uint32_t +static inline uint32_t z32_unorm_to_z24_unorm(uint32_t z) { /* z * 0xffffff / 0xffffffff */ return z >> 8; } -static INLINE uint32_t +static inline uint32_t z24_unorm_to_z32_unorm(uint32_t z) { /* z * 0xffffffff / 0xffffff */ @@ -68,42 +68,42 @@ z24_unorm_to_z32_unorm(uint32_t z) * z32_float conversion functions */ -static INLINE uint16_t +static inline uint16_t z32_float_to_z16_unorm(float z) { const float scale = 0xffff; return (uint16_t)(z * scale + 0.5f); } -static INLINE float +static inline float z16_unorm_to_z32_float(uint16_t z) { const float scale = 1.0 / 0xffff; return (float)(z * scale); } -static INLINE uint32_t +static inline uint32_t z32_float_to_z24_unorm(float z) { const double scale = 0xffffff; return (uint32_t)(z * scale) & 0xffffff; } -static INLINE float +static inline float z24_unorm_to_z32_float(uint32_t z) { const double scale = 1.0 / 0xffffff; return (float)(z * scale); } -static INLINE uint32_t +static inline uint32_t z32_float_to_z32_unorm(float z) { const double scale = 0xffffffff; return (uint32_t)(z * scale); } -static INLINE float +static inline float z32_unorm_to_z32_float(uint32_t z) { const double scale = 1.0 / 0xffffffff; diff --git a/src/gallium/auxiliary/util/u_half.h b/src/gallium/auxiliary/util/u_half.h index d340b9a7aef..d28fae3c77d 100644 --- a/src/gallium/auxiliary/util/u_half.h +++ b/src/gallium/auxiliary/util/u_half.h @@ -43,7 +43,7 @@ extern "C" { * https://gist.github.com/2144712 */ -static INLINE uint16_t +static inline uint16_t util_float_to_half(float f) { uint32_t sign_mask = 0x80000000; @@ -96,7 +96,7 @@ util_float_to_half(float f) return f16; } -static INLINE float +static inline float util_half_to_float(uint16_t f16) { union fi infnan; diff --git a/src/gallium/auxiliary/util/u_handle_table.c b/src/gallium/auxiliary/util/u_handle_table.c index 85302f1e194..42c4e44b644 100644 --- a/src/gallium/auxiliary/util/u_handle_table.c +++ b/src/gallium/auxiliary/util/u_handle_table.c @@ -96,7 +96,7 @@ handle_table_set_destroy(struct handle_table *ht, /** * Resize the table if necessary */ -static INLINE int +static inline int handle_table_resize(struct handle_table *ht, unsigned minimum_size) { @@ -126,7 +126,7 @@ handle_table_resize(struct handle_table *ht, } -static INLINE void +static inline void handle_table_clear(struct handle_table *ht, unsigned index) { diff --git a/src/gallium/auxiliary/util/u_hash_table.c b/src/gallium/auxiliary/util/u_hash_table.c index 06c8b5c91a5..a505fbc4d83 100644 --- a/src/gallium/auxiliary/util/u_hash_table.c +++ b/src/gallium/auxiliary/util/u_hash_table.c @@ -68,7 +68,7 @@ struct util_hash_table_item }; -static INLINE struct util_hash_table_item * +static inline struct util_hash_table_item * util_hash_table_item(struct cso_hash_iter iter) { return (struct util_hash_table_item *)cso_hash_iter_data(iter); @@ -98,7 +98,7 @@ util_hash_table_create(unsigned (*hash)(void *key), } -static INLINE struct cso_hash_iter +static inline struct cso_hash_iter util_hash_table_find_iter(struct util_hash_table *ht, void *key, unsigned key_hash) @@ -118,7 +118,7 @@ util_hash_table_find_iter(struct util_hash_table *ht, } -static INLINE struct util_hash_table_item * +static inline struct util_hash_table_item * util_hash_table_find_item(struct util_hash_table *ht, void *key, unsigned key_hash) diff --git a/src/gallium/auxiliary/util/u_inlines.h b/src/gallium/auxiliary/util/u_inlines.h index 95401621ec3..bb99a02ce49 100644 --- a/src/gallium/auxiliary/util/u_inlines.h +++ b/src/gallium/auxiliary/util/u_inlines.h @@ -51,13 +51,13 @@ extern "C" { */ -static INLINE void +static inline void pipe_reference_init(struct pipe_reference *reference, unsigned count) { p_atomic_set(&reference->count, count); } -static INLINE boolean +static inline boolean pipe_is_referenced(struct pipe_reference *reference) { return p_atomic_read(&reference->count) != 0; @@ -69,7 +69,7 @@ pipe_is_referenced(struct pipe_reference *reference) * Both 'ptr' and 'reference' may be NULL. * \return TRUE if the object's refcount hits zero and should be destroyed. */ -static INLINE boolean +static inline boolean pipe_reference_described(struct pipe_reference *ptr, struct pipe_reference *reference, debug_reference_descriptor get_desc) @@ -96,14 +96,14 @@ pipe_reference_described(struct pipe_reference *ptr, return destroy; } -static INLINE boolean +static inline boolean pipe_reference(struct pipe_reference *ptr, struct pipe_reference *reference) { return pipe_reference_described(ptr, reference, (debug_reference_descriptor)debug_describe_reference); } -static INLINE void +static inline void pipe_surface_reference(struct pipe_surface **ptr, struct pipe_surface *surf) { struct pipe_surface *old_surf = *ptr; @@ -120,7 +120,7 @@ pipe_surface_reference(struct pipe_surface **ptr, struct pipe_surface *surf) * of using a deleted context's surface_destroy() method when freeing a surface * that's shared by multiple contexts. */ -static INLINE void +static inline void pipe_surface_release(struct pipe_context *pipe, struct pipe_surface **ptr) { if (pipe_reference_described(&(*ptr)->reference, NULL, @@ -130,7 +130,7 @@ pipe_surface_release(struct pipe_context *pipe, struct pipe_surface **ptr) } -static INLINE void +static inline void pipe_resource_reference(struct pipe_resource **ptr, struct pipe_resource *tex) { struct pipe_resource *old_tex = *ptr; @@ -141,7 +141,7 @@ pipe_resource_reference(struct pipe_resource **ptr, struct pipe_resource *tex) *ptr = tex; } -static INLINE void +static inline void pipe_sampler_view_reference(struct pipe_sampler_view **ptr, struct pipe_sampler_view *view) { struct pipe_sampler_view *old_view = *ptr; @@ -158,7 +158,7 @@ pipe_sampler_view_reference(struct pipe_sampler_view **ptr, struct pipe_sampler_ * work-around for fixing a dangling context pointer problem when textures * are shared by multiple contexts. XXX fix this someday. */ -static INLINE void +static inline void pipe_sampler_view_release(struct pipe_context *ctx, struct pipe_sampler_view **ptr) { @@ -173,8 +173,18 @@ pipe_sampler_view_release(struct pipe_context *ctx, *ptr = NULL; } +static inline void +pipe_image_view_reference(struct pipe_image_view **ptr, struct pipe_image_view *view) +{ + struct pipe_image_view *old_view = *ptr; + + if (pipe_reference_described(&(*ptr)->reference, &view->reference, + (debug_reference_descriptor)debug_describe_image_view)) + old_view->context->image_view_destroy(old_view->context, old_view); + *ptr = view; +} -static INLINE void +static inline void pipe_so_target_reference(struct pipe_stream_output_target **ptr, struct pipe_stream_output_target *target) { @@ -186,7 +196,7 @@ pipe_so_target_reference(struct pipe_stream_output_target **ptr, *ptr = target; } -static INLINE void +static inline void pipe_surface_reset(struct pipe_context *ctx, struct pipe_surface* ps, struct pipe_resource *pt, unsigned level, unsigned layer) { @@ -199,7 +209,7 @@ pipe_surface_reset(struct pipe_context *ctx, struct pipe_surface* ps, ps->context = ctx; } -static INLINE void +static inline void pipe_surface_init(struct pipe_context *ctx, struct pipe_surface* ps, struct pipe_resource *pt, unsigned level, unsigned layer) { @@ -209,7 +219,7 @@ pipe_surface_init(struct pipe_context *ctx, struct pipe_surface* ps, } /* Return true if the surfaces are equal. */ -static INLINE boolean +static inline boolean pipe_surface_equal(struct pipe_surface *s1, struct pipe_surface *s2) { return s1->texture == s2->texture && @@ -233,7 +243,7 @@ pipe_surface_equal(struct pipe_surface *s1, struct pipe_surface *s2) * \param bind bitmask of PIPE_BIND_x flags * \param usage bitmask of PIPE_USAGE_x flags */ -static INLINE struct pipe_resource * +static inline struct pipe_resource * pipe_buffer_create( struct pipe_screen *screen, unsigned bind, unsigned usage, @@ -261,7 +271,7 @@ pipe_buffer_create( struct pipe_screen *screen, * \param access bitmask of PIPE_TRANSFER_x flags * \param transfer returns a transfer object */ -static INLINE void * +static inline void * pipe_buffer_map_range(struct pipe_context *pipe, struct pipe_resource *buffer, unsigned offset, @@ -292,7 +302,7 @@ pipe_buffer_map_range(struct pipe_context *pipe, * \param access bitmask of PIPE_TRANSFER_x flags * \param transfer returns a transfer object */ -static INLINE void * +static inline void * pipe_buffer_map(struct pipe_context *pipe, struct pipe_resource *buffer, unsigned access, @@ -302,14 +312,14 @@ pipe_buffer_map(struct pipe_context *pipe, } -static INLINE void +static inline void pipe_buffer_unmap(struct pipe_context *pipe, struct pipe_transfer *transfer) { pipe->transfer_unmap(pipe, transfer); } -static INLINE void +static inline void pipe_buffer_flush_mapped_range(struct pipe_context *pipe, struct pipe_transfer *transfer, unsigned offset, @@ -333,7 +343,7 @@ pipe_buffer_flush_mapped_range(struct pipe_context *pipe, pipe->transfer_flush_region(pipe, transfer, &box); } -static INLINE void +static inline void pipe_buffer_write(struct pipe_context *pipe, struct pipe_resource *buf, unsigned offset, @@ -367,7 +377,7 @@ pipe_buffer_write(struct pipe_context *pipe, * We can avoid GPU/CPU synchronization when writing range that has never * been written before. */ -static INLINE void +static inline void pipe_buffer_write_nooverlap(struct pipe_context *pipe, struct pipe_resource *buf, unsigned offset, unsigned size, @@ -393,7 +403,7 @@ pipe_buffer_write_nooverlap(struct pipe_context *pipe, * \param bind bitmask of PIPE_BIND_x flags * \param usage bitmask of PIPE_USAGE_x flags */ -static INLINE struct pipe_resource * +static inline struct pipe_resource * pipe_buffer_create_with_data(struct pipe_context *pipe, unsigned bind, unsigned usage, @@ -406,7 +416,7 @@ pipe_buffer_create_with_data(struct pipe_context *pipe, return res; } -static INLINE void +static inline void pipe_buffer_read(struct pipe_context *pipe, struct pipe_resource *buf, unsigned offset, @@ -433,7 +443,7 @@ pipe_buffer_read(struct pipe_context *pipe, * Map a resource for reading/writing. * \param access bitmask of PIPE_TRANSFER_x flags */ -static INLINE void * +static inline void * pipe_transfer_map(struct pipe_context *context, struct pipe_resource *resource, unsigned level, unsigned layer, @@ -456,7 +466,7 @@ pipe_transfer_map(struct pipe_context *context, * Map a 3D (texture) resource for reading/writing. * \param access bitmask of PIPE_TRANSFER_x flags */ -static INLINE void * +static inline void * pipe_transfer_map_3d(struct pipe_context *context, struct pipe_resource *resource, unsigned level, @@ -474,14 +484,14 @@ pipe_transfer_map_3d(struct pipe_context *context, &box, transfer); } -static INLINE void +static inline void pipe_transfer_unmap( struct pipe_context *context, struct pipe_transfer *transfer ) { context->transfer_unmap( context, transfer ); } -static INLINE void +static inline void pipe_set_constant_buffer(struct pipe_context *pipe, uint shader, uint index, struct pipe_resource *buf) { @@ -502,7 +512,7 @@ pipe_set_constant_buffer(struct pipe_context *pipe, uint shader, uint index, * Get the polygon offset enable/disable flag for the given polygon fill mode. * \param fill_mode one of PIPE_POLYGON_MODE_POINT/LINE/FILL */ -static INLINE boolean +static inline boolean util_get_offset(const struct pipe_rasterizer_state *templ, unsigned fill_mode) { @@ -519,7 +529,7 @@ util_get_offset(const struct pipe_rasterizer_state *templ, } } -static INLINE float +static inline float util_get_min_point_size(const struct pipe_rasterizer_state *state) { /* The point size should be clamped to this value at the rasterizer stage. @@ -529,7 +539,7 @@ util_get_min_point_size(const struct pipe_rasterizer_state *state) !state->multisample ? 1.0f : 0.0f; } -static INLINE void +static inline void util_query_clear_result(union pipe_query_result *result, unsigned type) { switch (type) { @@ -560,7 +570,7 @@ util_query_clear_result(union pipe_query_result *result, unsigned type) } /** Convert PIPE_TEXTURE_x to TGSI_TEXTURE_x */ -static INLINE unsigned +static inline unsigned util_pipe_tex_to_tgsi_tex(enum pipe_texture_target pipe_tex_target, unsigned nr_samples) { @@ -605,7 +615,7 @@ util_pipe_tex_to_tgsi_tex(enum pipe_texture_target pipe_tex_target, } -static INLINE void +static inline void util_copy_constant_buffer(struct pipe_constant_buffer *dst, const struct pipe_constant_buffer *src) { @@ -623,7 +633,7 @@ util_copy_constant_buffer(struct pipe_constant_buffer *dst, } } -static INLINE unsigned +static inline unsigned util_max_layer(const struct pipe_resource *r, unsigned level) { switch (r->target) { diff --git a/src/gallium/auxiliary/util/u_keymap.c b/src/gallium/auxiliary/util/u_keymap.c index ae14eda3cec..daa2991ced6 100644 --- a/src/gallium/auxiliary/util/u_keymap.c +++ b/src/gallium/auxiliary/util/u_keymap.c @@ -71,7 +71,7 @@ default_delete_func(const struct keymap *map, } -static INLINE struct keymap_item * +static inline struct keymap_item * hash_table_item(struct cso_hash_iter iter) { return (struct keymap_item *) cso_hash_iter_data(iter); @@ -143,7 +143,7 @@ util_delete_keymap(struct keymap *map, void *user) } -static INLINE struct cso_hash_iter +static inline struct cso_hash_iter hash_table_find_iter(const struct keymap *map, const void *key, unsigned key_hash) { @@ -162,7 +162,7 @@ hash_table_find_iter(const struct keymap *map, const void *key, } -static INLINE struct keymap_item * +static inline struct keymap_item * hash_table_find_item(const struct keymap *map, const void *key, unsigned key_hash) { diff --git a/src/gallium/auxiliary/util/u_linear.h b/src/gallium/auxiliary/util/u_linear.h index 81ffc9fb27d..87e52a344d4 100644 --- a/src/gallium/auxiliary/util/u_linear.h +++ b/src/gallium/auxiliary/util/u_linear.h @@ -89,7 +89,7 @@ void pipe_linear_fill_info(struct pipe_tile_info *t, unsigned tile_width, unsigned tile_height, unsigned tiles_x, unsigned tiles_y); -static INLINE boolean pipe_linear_check_tile(const struct pipe_tile_info *t) +static inline boolean pipe_linear_check_tile(const struct pipe_tile_info *t) { if (t->tile.size != t->block.size * t->cols * t->rows) return FALSE; diff --git a/src/gallium/auxiliary/util/u_math.c b/src/gallium/auxiliary/util/u_math.c index ae9e9513b04..c58af911be7 100644 --- a/src/gallium/auxiliary/util/u_math.c +++ b/src/gallium/auxiliary/util/u_math.c @@ -48,7 +48,7 @@ init_pow2_table(void) { int i; for (i = 0; i < POW2_TABLE_SIZE; i++) - pow2_table[i] = (float) pow(2.0, (i - POW2_TABLE_OFFSET) / POW2_TABLE_SCALE); + pow2_table[i] = exp2f((i - POW2_TABLE_OFFSET) / POW2_TABLE_SCALE); } diff --git a/src/gallium/auxiliary/util/u_math.h b/src/gallium/auxiliary/util/u_math.h index 3b4040f0ee2..56bd185f527 100644 --- a/src/gallium/auxiliary/util/u_math.h +++ b/src/gallium/auxiliary/util/u_math.h @@ -92,7 +92,7 @@ union di { /** * Extract the IEEE float32 exponent. */ -static INLINE signed +static inline signed util_get_float32_exponent(float x) { union fi f; @@ -112,7 +112,7 @@ util_get_float32_exponent(float x) * Compute exp2(ipart) with i << ipart * Compute exp2(fpart) with lookup table. */ -static INLINE float +static inline float util_fast_exp2(float x) { int32_t ipart; @@ -143,7 +143,7 @@ util_fast_exp2(float x) /** * Fast approximation to exp(x). */ -static INLINE float +static inline float util_fast_exp(float x) { const float k = 1.44269f; /* = log2(e) */ @@ -160,7 +160,7 @@ extern float log2_table[LOG2_TABLE_SIZE]; /** * Fast approximation to log2(x). */ -static INLINE float +static inline float util_fast_log2(float x) { union fi num; @@ -176,7 +176,7 @@ util_fast_log2(float x) /** * Fast approximation to x^y. */ -static INLINE float +static inline float util_fast_pow(float x, float y) { return util_fast_exp2(util_fast_log2(x) * y); @@ -184,7 +184,7 @@ util_fast_pow(float x, float y) /* Note that this counts zero as a power of two. */ -static INLINE boolean +static inline boolean util_is_power_of_two( unsigned v ) { return (v & (v-1)) == 0; @@ -194,7 +194,7 @@ util_is_power_of_two( unsigned v ) /** * Floor(x), returned as int. */ -static INLINE int +static inline int util_ifloor(float f) { int ai, bi; @@ -211,7 +211,7 @@ util_ifloor(float f) /** * Round float to nearest int. */ -static INLINE int +static inline int util_iround(float f) { #if defined(PIPE_CC_GCC) && defined(PIPE_ARCH_X86) @@ -237,10 +237,10 @@ util_iround(float f) /** * Approximate floating point comparison */ -static INLINE boolean +static inline boolean util_is_approx(float a, float b, float tol) { - return fabs(b - a) <= tol; + return fabsf(b - a) <= tol; } @@ -256,7 +256,7 @@ util_is_approx(float a, float b, float tol) /** * Single-float */ -static INLINE boolean +static inline boolean util_is_inf_or_nan(float x) { union fi tmp; @@ -265,7 +265,7 @@ util_is_inf_or_nan(float x) } -static INLINE boolean +static inline boolean util_is_nan(float x) { union fi tmp; @@ -274,7 +274,7 @@ util_is_nan(float x) } -static INLINE int +static inline int util_inf_sign(float x) { union fi tmp; @@ -290,7 +290,7 @@ util_inf_sign(float x) /** * Double-float */ -static INLINE boolean +static inline boolean util_is_double_inf_or_nan(double x) { union di tmp; @@ -299,7 +299,7 @@ util_is_double_inf_or_nan(double x) } -static INLINE boolean +static inline boolean util_is_double_nan(double x) { union di tmp; @@ -308,7 +308,7 @@ util_is_double_nan(double x) } -static INLINE int +static inline int util_double_inf_sign(double x) { union di tmp; @@ -324,21 +324,21 @@ util_double_inf_sign(double x) /** * Half-float */ -static INLINE boolean +static inline boolean util_is_half_inf_or_nan(int16_t x) { return (x & 0x7c00) == 0x7c00; } -static INLINE boolean +static inline boolean util_is_half_nan(int16_t x) { return (x & 0x7fff) > 0x7c00; } -static INLINE int +static inline int util_half_inf_sign(int16_t x) { if ((x & 0x7fff) != 0x7c00) { @@ -359,7 +359,7 @@ util_half_inf_sign(int16_t x) #if defined(_MSC_VER) && (_M_IX86 || _M_AMD64 || _M_IA64) unsigned char _BitScanForward(unsigned long* Index, unsigned long Mask); #pragma intrinsic(_BitScanForward) -static INLINE +static inline unsigned long ffs( unsigned long u ) { unsigned long i; @@ -369,7 +369,7 @@ unsigned long ffs( unsigned long u ) return 0; } #elif defined(PIPE_CC_MSVC) && defined(PIPE_ARCH_X86) -static INLINE +static inline unsigned ffs( unsigned u ) { unsigned i; @@ -409,7 +409,7 @@ unsigned ffs( unsigned u ) * Find last bit set in a word. The least significant bit is 1. * Return 0 if no bits are set. */ -static INLINE unsigned +static inline unsigned util_last_bit(unsigned u) { #if defined(HAVE___BUILTIN_CLZ) @@ -428,7 +428,7 @@ util_last_bit(unsigned u) * Find last bit set in a word. The least significant bit is 1. * Return 0 if no bits are set. */ -static INLINE unsigned +static inline unsigned util_last_bit64(uint64_t u) { #if defined(HAVE___BUILTIN_CLZLL) @@ -448,7 +448,7 @@ util_last_bit64(uint64_t u) * significant bit is 1. * Return 0 if no bits are set. */ -static INLINE unsigned +static inline unsigned util_last_bit_signed(int i) { if (i >= 0) @@ -465,7 +465,7 @@ util_last_bit_signed(int i) * } * */ -static INLINE int +static inline int u_bit_scan(unsigned *mask) { int i = ffs(*mask) - 1; @@ -474,7 +474,7 @@ u_bit_scan(unsigned *mask) } #ifndef _MSC_VER -static INLINE int +static inline int u_bit_scan64(uint64_t *mask) { int i = ffsll(*mask) - 1; @@ -486,7 +486,7 @@ u_bit_scan64(uint64_t *mask) /** * Return float bits. */ -static INLINE unsigned +static inline unsigned fui( float f ) { union fi fi; @@ -494,7 +494,7 @@ fui( float f ) return fi.ui; } -static INLINE float +static inline float uif(uint32_t ui) { union fi fi; @@ -507,7 +507,7 @@ uif(uint32_t ui) * Convert ubyte to float in [0, 1]. * XXX a 256-entry lookup table would be slightly faster. */ -static INLINE float +static inline float ubyte_to_float(ubyte ub) { return (float) ub * (1.0f / 255.0f); @@ -517,7 +517,7 @@ ubyte_to_float(ubyte ub) /** * Convert float in [0,1] to ubyte in [0,255] with clamping. */ -static INLINE ubyte +static inline ubyte float_to_ubyte(float f) { union fi tmp; @@ -535,13 +535,13 @@ float_to_ubyte(float f) } } -static INLINE float +static inline float byte_to_float_tex(int8_t b) { return (b == -128) ? -1.0F : b * 1.0F / 127.0F; } -static INLINE int8_t +static inline int8_t float_to_byte_tex(float f) { return (int8_t) (127.0F * f); @@ -550,7 +550,7 @@ float_to_byte_tex(float f) /** * Calc log base 2 */ -static INLINE unsigned +static inline unsigned util_logbase2(unsigned n) { #if defined(HAVE___BUILTIN_CLZ) @@ -570,7 +570,7 @@ util_logbase2(unsigned n) /** * Returns the smallest power of two >= x */ -static INLINE unsigned +static inline unsigned util_next_power_of_two(unsigned x) { #if defined(HAVE___BUILTIN_CLZ) @@ -602,7 +602,7 @@ util_next_power_of_two(unsigned x) /** * Return number of bits set in n. */ -static INLINE unsigned +static inline unsigned util_bitcount(unsigned n) { #if defined(HAVE___BUILTIN_POPCOUNT) @@ -623,7 +623,7 @@ util_bitcount(unsigned n) } -static INLINE unsigned +static inline unsigned util_bitcount64(uint64_t n) { #ifdef HAVE___BUILTIN_POPCOUNTLL @@ -639,7 +639,7 @@ util_bitcount64(uint64_t n) * Algorithm taken from: * http://stackoverflow.com/questions/9144800/c-reverse-bits-in-unsigned-integer */ -static INLINE unsigned +static inline unsigned util_bitreverse(unsigned n) { n = ((n >> 1) & 0x55555555u) | ((n & 0x55555555u) << 1); @@ -671,7 +671,7 @@ util_bitreverse(unsigned n) /** * Reverse byte order of a 32 bit word. */ -static INLINE uint32_t +static inline uint32_t util_bswap32(uint32_t n) { #if defined(HAVE___BUILTIN_BSWAP32) @@ -687,7 +687,7 @@ util_bswap32(uint32_t n) /** * Reverse byte order of a 64bit word. */ -static INLINE uint64_t +static inline uint64_t util_bswap64(uint64_t n) { #if defined(HAVE___BUILTIN_BSWAP64) @@ -702,14 +702,14 @@ util_bswap64(uint64_t n) /** * Reverse byte order of a 16 bit word. */ -static INLINE uint16_t +static inline uint16_t util_bswap16(uint16_t n) { return (n >> 8) | (n << 8); } -static INLINE void* +static inline void* util_memcpy_cpu_to_le32(void * restrict dest, const void * restrict src, size_t n) { #ifdef PIPE_ARCH_BIG_ENDIAN @@ -746,7 +746,7 @@ util_memcpy_cpu_to_le32(void * restrict dest, const void * restrict src, size_t /** * Align a value, only works pot alignemnts. */ -static INLINE int +static inline int align(int value, int alignment) { return (value + alignment - 1) & ~(alignment - 1); @@ -755,7 +755,7 @@ align(int value, int alignment) /** * Works like align but on npot alignments. */ -static INLINE size_t +static inline size_t util_align_npot(size_t value, size_t alignment) { if (value % alignment) @@ -763,7 +763,7 @@ util_align_npot(size_t value, size_t alignment) return value; } -static INLINE unsigned +static inline unsigned u_minify(unsigned value, unsigned levels) { return MAX2(1, value >> levels); @@ -796,13 +796,13 @@ do { \ #endif -static INLINE uint32_t +static inline uint32_t util_unsigned_fixed(float value, unsigned frac_bits) { return value < 0 ? 0 : (uint32_t)(value * (1<<frac_bits)); } -static INLINE int32_t +static inline int32_t util_signed_fixed(float value, unsigned frac_bits) { return (int32_t)(value * (1<<frac_bits)); diff --git a/src/gallium/auxiliary/util/u_memory.h b/src/gallium/auxiliary/util/u_memory.h index 9ff6c7da919..7fe0fe6f053 100644 --- a/src/gallium/auxiliary/util/u_memory.h +++ b/src/gallium/auxiliary/util/u_memory.h @@ -67,7 +67,7 @@ extern "C" { /** * Duplicate a block of memory. */ -static INLINE void * +static inline void * mem_dup(const void *src, uint size) { void *dup = MALLOC(size); diff --git a/src/gallium/auxiliary/util/u_mm.c b/src/gallium/auxiliary/util/u_mm.c index 82f83702d1e..2069b56f464 100644 --- a/src/gallium/auxiliary/util/u_mm.c +++ b/src/gallium/auxiliary/util/u_mm.c @@ -224,7 +224,7 @@ u_mmFindBlock(struct mem_block *heap, int start) } -static INLINE int +static inline int Join2Blocks(struct mem_block *p) { /* XXX there should be some assertions here */ diff --git a/src/gallium/auxiliary/util/u_pack_color.h b/src/gallium/auxiliary/util/u_pack_color.h index e0c9018f8ef..b882502b7ba 100644 --- a/src/gallium/auxiliary/util/u_pack_color.h +++ b/src/gallium/auxiliary/util/u_pack_color.h @@ -60,7 +60,7 @@ union util_color { /** * Pack ubyte R,G,B,A into dest pixel. */ -static INLINE void +static inline void util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a, enum pipe_format format, union util_color *uc) { @@ -161,7 +161,7 @@ util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a, /** * Unpack RGBA from a packed pixel, returning values as ubytes in [0,255]. */ -static INLINE void +static inline void util_unpack_color_ub(enum pipe_format format, union util_color *uc, ubyte *r, ubyte *g, ubyte *b, ubyte *a) { @@ -333,7 +333,7 @@ util_unpack_color_ub(enum pipe_format format, union util_color *uc, * This will not work (and might not really be useful with float input) * for pure integer formats (which lack the pack_rgba_float function). */ -static INLINE void +static inline void util_pack_color(const float rgba[4], enum pipe_format format, union util_color *uc) { ubyte r = 0; @@ -437,7 +437,7 @@ util_pack_color(const float rgba[4], enum pipe_format format, union util_color * /* Integer versions of util_pack_z and util_pack_z_stencil - useful for * constructing clear masks. */ -static INLINE uint32_t +static inline uint32_t util_pack_mask_z(enum pipe_format format, uint32_t z) { switch (format) { @@ -462,7 +462,7 @@ util_pack_mask_z(enum pipe_format format, uint32_t z) } -static INLINE uint64_t +static inline uint64_t util_pack64_mask_z(enum pipe_format format, uint32_t z) { switch (format) { @@ -474,7 +474,7 @@ util_pack64_mask_z(enum pipe_format format, uint32_t z) } -static INLINE uint32_t +static inline uint32_t util_pack_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s) { uint32_t packed = util_pack_mask_z(format, z); @@ -497,7 +497,7 @@ util_pack_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s) } -static INLINE uint64_t +static inline uint64_t util_pack64_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s) { uint64_t packed; @@ -516,7 +516,7 @@ util_pack64_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s) /** * Note: it's assumed that z is in [0,1] */ -static INLINE uint32_t +static inline uint32_t util_pack_z(enum pipe_format format, double z) { union fi fui; @@ -558,7 +558,7 @@ util_pack_z(enum pipe_format format, double z) } -static INLINE uint64_t +static inline uint64_t util_pack64_z(enum pipe_format format, double z) { union fi fui; @@ -580,7 +580,7 @@ util_pack64_z(enum pipe_format format, double z) * Pack Z and/or stencil values into a 32-bit value described by format. * Note: it's assumed that z is in [0,1] and s in [0,255] */ -static INLINE uint32_t +static inline uint32_t util_pack_z_stencil(enum pipe_format format, double z, uint8_t s) { uint32_t packed = util_pack_z(format, z); @@ -603,7 +603,7 @@ util_pack_z_stencil(enum pipe_format format, double z, uint8_t s) } -static INLINE uint64_t +static inline uint64_t util_pack64_z_stencil(enum pipe_format format, double z, uint8_t s) { uint64_t packed; @@ -624,7 +624,7 @@ util_pack64_z_stencil(enum pipe_format format, double z, uint8_t s) /** * Pack 4 ubytes into a 4-byte word */ -static INLINE unsigned +static inline unsigned pack_ub4(ubyte b0, ubyte b1, ubyte b2, ubyte b3) { return ((((unsigned int)b0) << 0) | @@ -637,7 +637,7 @@ pack_ub4(ubyte b0, ubyte b1, ubyte b2, ubyte b3) /** * Pack/convert 4 floats into one 4-byte word. */ -static INLINE unsigned +static inline unsigned pack_ui32_float4(float a, float b, float c, float d) { return pack_ub4( float_to_ubyte(a), diff --git a/src/gallium/auxiliary/util/u_pointer.h b/src/gallium/auxiliary/util/u_pointer.h index 30c23b79831..4f7a27ca61d 100644 --- a/src/gallium/auxiliary/util/u_pointer.h +++ b/src/gallium/auxiliary/util/u_pointer.h @@ -34,7 +34,7 @@ extern "C" { #endif -static INLINE intptr_t +static inline intptr_t pointer_to_intptr( const void *p ) { union { @@ -45,7 +45,7 @@ pointer_to_intptr( const void *p ) return pi.i; } -static INLINE void * +static inline void * intptr_to_pointer( intptr_t i ) { union { @@ -56,7 +56,7 @@ intptr_to_pointer( intptr_t i ) return pi.p; } -static INLINE uintptr_t +static inline uintptr_t pointer_to_uintptr( const void *ptr ) { union { @@ -67,7 +67,7 @@ pointer_to_uintptr( const void *ptr ) return pu.u; } -static INLINE void * +static inline void * uintptr_to_pointer( uintptr_t u ) { union { @@ -81,7 +81,7 @@ uintptr_to_pointer( uintptr_t u ) /** * Return a pointer aligned to next multiple of N bytes. */ -static INLINE void * +static inline void * align_pointer( const void *unaligned, uintptr_t alignment ) { uintptr_t aligned = (pointer_to_uintptr( unaligned ) + alignment - 1) & ~(alignment - 1); @@ -92,7 +92,7 @@ align_pointer( const void *unaligned, uintptr_t alignment ) /** * Return a pointer aligned to next multiple of 16 bytes. */ -static INLINE void * +static inline void * align16( void *unaligned ) { return align_pointer( unaligned, 16 ); @@ -100,7 +100,7 @@ align16( void *unaligned ) typedef void (*func_pointer)(void); -static INLINE func_pointer +static inline func_pointer pointer_to_func( void *p ) { union { @@ -111,7 +111,7 @@ pointer_to_func( void *p ) return pf.f; } -static INLINE void * +static inline void * func_to_pointer( func_pointer f ) { union { diff --git a/src/gallium/auxiliary/util/u_prim.h b/src/gallium/auxiliary/util/u_prim.h index b2dd44df230..366801545ed 100644 --- a/src/gallium/auxiliary/util/u_prim.h +++ b/src/gallium/auxiliary/util/u_prim.h @@ -46,7 +46,7 @@ struct u_prim_vertex_count { * Decompose a primitive that is a loop, a strip, or a fan. Return the * original primitive if it is already decomposed. */ -static INLINE unsigned +static inline unsigned u_decomposed_prim(unsigned prim) { switch (prim) { @@ -71,7 +71,7 @@ u_decomposed_prim(unsigned prim) * Reduce a primitive to one of PIPE_PRIM_POINTS, PIPE_PRIM_LINES, and * PIPE_PRIM_TRIANGLES. */ -static INLINE unsigned +static inline unsigned u_reduced_prim(unsigned prim) { switch (prim) { @@ -91,7 +91,7 @@ u_reduced_prim(unsigned prim) /** * Re-assemble a primitive to remove its adjacency. */ -static INLINE unsigned +static inline unsigned u_assembled_prim(unsigned prim) { switch (prim) { @@ -113,7 +113,7 @@ u_assembled_prim(unsigned prim) * source file, it will increase the size of the binary slightly more than * expected because of the use of a table. */ -static INLINE const struct u_prim_vertex_count * +static inline const struct u_prim_vertex_count * u_prim_vertex_count(unsigned prim) { static const struct u_prim_vertex_count prim_table[PIPE_PRIM_MAX] = { @@ -140,7 +140,7 @@ u_prim_vertex_count(unsigned prim) * Given a vertex count, return the number of primitives. * For polygons, return the number of triangles. */ -static INLINE unsigned +static inline unsigned u_prims_for_vertices(unsigned prim, unsigned num) { const struct u_prim_vertex_count *info = u_prim_vertex_count(prim); @@ -151,7 +151,7 @@ u_prims_for_vertices(unsigned prim, unsigned num) return 1 + ((num - info->min) / info->incr); } -static INLINE boolean u_validate_pipe_prim( unsigned pipe_prim, unsigned nr ) +static inline boolean u_validate_pipe_prim( unsigned pipe_prim, unsigned nr ) { const struct u_prim_vertex_count *count = u_prim_vertex_count(pipe_prim); @@ -159,7 +159,7 @@ static INLINE boolean u_validate_pipe_prim( unsigned pipe_prim, unsigned nr ) } -static INLINE boolean u_trim_pipe_prim( unsigned pipe_prim, unsigned *nr ) +static inline boolean u_trim_pipe_prim( unsigned pipe_prim, unsigned *nr ) { const struct u_prim_vertex_count *count = u_prim_vertex_count(pipe_prim); @@ -174,7 +174,7 @@ static INLINE boolean u_trim_pipe_prim( unsigned pipe_prim, unsigned *nr ) } } -static INLINE unsigned +static inline unsigned u_vertices_per_prim(int primitive) { switch(primitive) { @@ -216,7 +216,7 @@ u_vertices_per_prim(int primitive) * statistics depend on knowing the exact number of decomposed * primitives for a set of vertices. */ -static INLINE unsigned +static inline unsigned u_decomposed_prims_for_vertices(int primitive, int vertices) { switch (primitive) { @@ -263,7 +263,7 @@ u_decomposed_prims_for_vertices(int primitive, int vertices) * count. Each quad is treated as two triangles. Polygons are treated as * triangle fans. */ -static INLINE unsigned +static inline unsigned u_reduced_prims_for_vertices(int primitive, int vertices) { switch (primitive) { diff --git a/src/gallium/auxiliary/util/u_range.h b/src/gallium/auxiliary/util/u_range.h index efe25ef5e42..a1da5e5a6f0 100644 --- a/src/gallium/auxiliary/util/u_range.h +++ b/src/gallium/auxiliary/util/u_range.h @@ -47,7 +47,7 @@ struct util_range { }; -static INLINE void +static inline void util_range_set_empty(struct util_range *range) { range->start = ~0; @@ -55,7 +55,7 @@ util_range_set_empty(struct util_range *range) } /* This is like a union of two sets. */ -static INLINE void +static inline void util_range_add(struct util_range *range, unsigned start, unsigned end) { if (start < range->start || end > range->end) { @@ -66,7 +66,7 @@ util_range_add(struct util_range *range, unsigned start, unsigned end) } } -static INLINE boolean +static inline boolean util_ranges_intersect(struct util_range *range, unsigned start, unsigned end) { return MAX2(start, range->start) < MIN2(end, range->end); @@ -75,14 +75,14 @@ util_ranges_intersect(struct util_range *range, unsigned start, unsigned end) /* Init/deinit */ -static INLINE void +static inline void util_range_init(struct util_range *range) { pipe_mutex_init(range->write_mutex); util_range_set_empty(range); } -static INLINE void +static inline void util_range_destroy(struct util_range *range) { pipe_mutex_destroy(range->write_mutex); diff --git a/src/gallium/auxiliary/util/u_rect.h b/src/gallium/auxiliary/util/u_rect.h index cf29dff0d02..b26f671f313 100644 --- a/src/gallium/auxiliary/util/u_rect.h +++ b/src/gallium/auxiliary/util/u_rect.h @@ -43,7 +43,7 @@ struct u_rect { /* Do two rectangles intersect? */ -static INLINE boolean +static inline boolean u_rect_test_intersection(const struct u_rect *a, const struct u_rect *b) { @@ -55,7 +55,7 @@ u_rect_test_intersection(const struct u_rect *a, /* Find the intersection of two rectangles known to intersect. */ -static INLINE void +static inline void u_rect_find_intersection(const struct u_rect *a, struct u_rect *b) { @@ -68,13 +68,13 @@ u_rect_find_intersection(const struct u_rect *a, } -static INLINE int +static inline int u_rect_area(const struct u_rect *r) { return (r->x1 - r->x0) * (r->y1 - r->y0); } -static INLINE void +static inline void u_rect_possible_intersection(const struct u_rect *a, struct u_rect *b) { @@ -88,7 +88,7 @@ u_rect_possible_intersection(const struct u_rect *a, /* Set @d to a rectangle that covers both @a and @b. */ -static INLINE void +static inline void u_rect_union(struct u_rect *d, const struct u_rect *a, const struct u_rect *b) { d->x0 = MIN2(a->x0, b->x0); diff --git a/src/gallium/auxiliary/util/u_resource.h b/src/gallium/auxiliary/util/u_resource.h index a5e091fd66e..6736476f4da 100644 --- a/src/gallium/auxiliary/util/u_resource.h +++ b/src/gallium/auxiliary/util/u_resource.h @@ -36,7 +36,7 @@ util_resource_size(const struct pipe_resource *res); * * Note that this function returns true for single-layered array textures. */ -static INLINE boolean +static inline boolean util_resource_is_array_texture(const struct pipe_resource *res) { switch (res->target) { diff --git a/src/gallium/auxiliary/util/u_ringbuffer.c b/src/gallium/auxiliary/util/u_ringbuffer.c index 648b105b137..5816b781660 100644 --- a/src/gallium/auxiliary/util/u_ringbuffer.c +++ b/src/gallium/auxiliary/util/u_ringbuffer.c @@ -56,7 +56,7 @@ void util_ringbuffer_destroy( struct util_ringbuffer *ring ) /** * Return number of free entries in the ring */ -static INLINE unsigned util_ringbuffer_space( const struct util_ringbuffer *ring ) +static inline unsigned util_ringbuffer_space( const struct util_ringbuffer *ring ) { return (ring->tail - (ring->head + 1)) & ring->mask; } @@ -64,7 +64,7 @@ static INLINE unsigned util_ringbuffer_space( const struct util_ringbuffer *ring /** * Is the ring buffer empty? */ -static INLINE boolean util_ringbuffer_empty( const struct util_ringbuffer *ring ) +static inline boolean util_ringbuffer_empty( const struct util_ringbuffer *ring ) { return util_ringbuffer_space(ring) == ring->mask; } diff --git a/src/gallium/auxiliary/util/u_snprintf.c b/src/gallium/auxiliary/util/u_snprintf.c deleted file mode 100644 index 39e9b70d0f8..00000000000 --- a/src/gallium/auxiliary/util/u_snprintf.c +++ /dev/null @@ -1,1480 +0,0 @@ -/* - * Copyright (c) 1995 Patrick Powell. - * - * This code is based on code written by Patrick Powell <[email protected]>. - * It may be used for any purpose as long as this notice remains intact on all - * source code distributions. - */ - -/* - * Copyright (c) 2008 Holger Weiss. - * - * This version of the code is maintained by Holger Weiss <[email protected]>. - * My changes to the code may freely be used, modified and/or redistributed for - * any purpose. It would be nice if additions and fixes to this file (including - * trivial code cleanups) would be sent back in order to let me include them in - * the version available at <http://www.jhweiss.de/software/snprintf.html>. - * However, this is not a requirement for using or redistributing (possibly - * modified) versions of this file, nor is leaving this notice intact mandatory. - */ - -/* - * History - * - * 2008-01-20 Holger Weiss <[email protected]> for C99-snprintf 1.1: - * - * Fixed the detection of infinite floating point values on IRIX (and - * possibly other systems) and applied another few minor cleanups. - * - * 2008-01-06 Holger Weiss <[email protected]> for C99-snprintf 1.0: - * - * Added a lot of new features, fixed many bugs, and incorporated various - * improvements done by Andrew Tridgell <[email protected]>, Russ Allbery - * <[email protected]>, Hrvoje Niksic <[email protected]>, Damien Miller - * <[email protected]>, and others for the Samba, INN, Wget, and OpenSSH - * projects. The additions include: support the "e", "E", "g", "G", and - * "F" conversion specifiers (and use conversion style "f" or "F" for the - * still unsupported "a" and "A" specifiers); support the "hh", "ll", "j", - * "t", and "z" length modifiers; support the "#" flag and the (non-C99) - * "'" flag; use localeconv(3) (if available) to get both the current - * locale's decimal point character and the separator between groups of - * digits; fix the handling of various corner cases of field width and - * precision specifications; fix various floating point conversion bugs; - * handle infinite and NaN floating point values; don't attempt to write to - * the output buffer (which may be NULL) if a size of zero was specified; - * check for integer overflow of the field width, precision, and return - * values and during the floating point conversion; use the OUTCHAR() macro - * instead of a function for better performance; provide asprintf(3) and - * vasprintf(3) functions; add new test cases. The replacement functions - * have been renamed to use an "rpl_" prefix, the function calls in the - * main project (and in this file) must be redefined accordingly for each - * replacement function which is needed (by using Autoconf or other means). - * Various other minor improvements have been applied and the coding style - * was cleaned up for consistency. - * - * 2007-07-23 Holger Weiss <[email protected]> for Mutt 1.5.13: - * - * C99 compliant snprintf(3) and vsnprintf(3) functions return the number - * of characters that would have been written to a sufficiently sized - * buffer (excluding the '\0'). The original code simply returned the - * length of the resulting output string, so that's been fixed. - * - * 1998-03-05 Michael Elkins <[email protected]> for Mutt 0.90.8: - * - * The original code assumed that both snprintf(3) and vsnprintf(3) were - * missing. Some systems only have snprintf(3) but not vsnprintf(3), so - * the code is now broken down under HAVE_SNPRINTF and HAVE_VSNPRINTF. - * - * 1998-01-27 Thomas Roessler <[email protected]> for Mutt 0.89i: - * - * The PGP code was using unsigned hexadecimal formats. Unfortunately, - * unsigned formats simply didn't work. - * - * 1997-10-22 Brandon Long <[email protected]> for Mutt 0.87.1: - * - * Ok, added some minimal floating point support, which means this probably - * requires libm on most operating systems. Don't yet support the exponent - * (e,E) and sigfig (g,G). Also, fmtint() was pretty badly broken, it just - * wasn't being exercised in ways which showed it, so that's been fixed. - * Also, formatted the code to Mutt conventions, and removed dead code left - * over from the original. Also, there is now a builtin-test, run with: - * gcc -DTEST_SNPRINTF -o snprintf snprintf.c -lm && ./snprintf - * - * 2996-09-15 Brandon Long <[email protected]> for Mutt 0.43: - * - * This was ugly. It is still ugly. I opted out of floating point - * numbers, but the formatter understands just about everything from the - * normal C string format, at least as far as I can tell from the Solaris - * 2.5 printf(3S) man page. - */ - -/* - * ToDo - * - * - Add wide character support. - * - Add support for "%a" and "%A" conversions. - * - Create test routines which predefine the expected results. Our test cases - * usually expose bugs in system implementations rather than in ours :-) - */ - -/* - * Usage - * - * 1) The following preprocessor macros should be defined to 1 if the feature or - * file in question is available on the target system (by using Autoconf or - * other means), though basic functionality should be available as long as - * HAVE_STDARG_H and HAVE_STDLIB_H are defined correctly: - * - * HAVE_VSNPRINTF - * HAVE_SNPRINTF - * HAVE_VASPRINTF - * HAVE_ASPRINTF - * HAVE_STDARG_H - * HAVE_STDDEF_H - * HAVE_STDINT_H - * HAVE_STDLIB_H - * HAVE_INTTYPES_H - * HAVE_LOCALE_H - * HAVE_LOCALECONV - * HAVE_LCONV_DECIMAL_POINT - * HAVE_LCONV_THOUSANDS_SEP - * HAVE_LONG_DOUBLE - * HAVE_LONG_LONG_INT - * HAVE_UNSIGNED_LONG_LONG_INT - * HAVE_INTMAX_T - * HAVE_UINTMAX_T - * HAVE_UINTPTR_T - * HAVE_PTRDIFF_T - * HAVE_VA_COPY - * HAVE___VA_COPY - * - * 2) The calls to the functions which should be replaced must be redefined - * throughout the project files (by using Autoconf or other means): - * - * #define vsnprintf rpl_vsnprintf - * #define snprintf rpl_snprintf - * #define vasprintf rpl_vasprintf - * #define asprintf rpl_asprintf - * - * 3) The required replacement functions should be declared in some header file - * included throughout the project files: - * - * #if HAVE_CONFIG_H - * #include <config.h> - * #endif - * #if HAVE_STDARG_H - * #include <stdarg.h> - * #if !HAVE_VSNPRINTF - * int rpl_vsnprintf(char *, size_t, const char *, va_list); - * #endif - * #if !HAVE_SNPRINTF - * int rpl_snprintf(char *, size_t, const char *, ...); - * #endif - * #if !HAVE_VASPRINTF - * int rpl_vasprintf(char **, const char *, va_list); - * #endif - * #if !HAVE_ASPRINTF - * int rpl_asprintf(char **, const char *, ...); - * #endif - * #endif - * - * Autoconf macros for handling step 1 and step 2 are available at - * <http://www.jhweiss.de/software/snprintf.html>. - */ - -#include "pipe/p_config.h" - -#if HAVE_CONFIG_H -#include <config.h> -#else -#ifdef _MSC_VER -#define vsnprintf util_vsnprintf -#define snprintf util_snprintf -#define HAVE_VSNPRINTF 0 -#define HAVE_SNPRINTF 0 -#define HAVE_VASPRINTF 1 /* not needed */ -#define HAVE_ASPRINTF 1 /* not needed */ -#define HAVE_STDARG_H 1 -#define HAVE_STDDEF_H 1 -#define HAVE_STDINT_H 1 -#define HAVE_STDLIB_H 1 -#define HAVE_INTTYPES_H 0 -#define HAVE_LOCALE_H 0 -#define HAVE_LOCALECONV 0 -#define HAVE_LCONV_DECIMAL_POINT 0 -#define HAVE_LCONV_THOUSANDS_SEP 0 -#define HAVE_LONG_DOUBLE 0 -#define HAVE_LONG_LONG_INT 1 -#define HAVE_UNSIGNED_LONG_LONG_INT 1 -#define HAVE_INTMAX_T 0 -#define HAVE_UINTMAX_T 0 -#define HAVE_UINTPTR_T 1 -#define HAVE_PTRDIFF_T 1 -#define HAVE_VA_COPY 0 -#define HAVE___VA_COPY 0 -#else -#define HAVE_VSNPRINTF 1 -#define HAVE_SNPRINTF 1 -#define HAVE_VASPRINTF 1 -#define HAVE_ASPRINTF 1 -#endif -#endif /* HAVE_CONFIG_H */ - -#if !HAVE_SNPRINTF || !HAVE_VSNPRINTF || !HAVE_ASPRINTF || !HAVE_VASPRINTF -#include <stdio.h> /* For NULL, size_t, vsnprintf(3), and vasprintf(3). */ -#ifdef VA_START -#undef VA_START -#endif /* defined(VA_START) */ -#ifdef VA_SHIFT -#undef VA_SHIFT -#endif /* defined(VA_SHIFT) */ -#if HAVE_STDARG_H -#include <stdarg.h> -#define VA_START(ap, last) va_start(ap, last) -#define VA_SHIFT(ap, value, type) /* No-op for ANSI C. */ -#else /* Assume <varargs.h> is available. */ -#include <varargs.h> -#define VA_START(ap, last) va_start(ap) /* "last" is ignored. */ -#define VA_SHIFT(ap, value, type) value = va_arg(ap, type) -#endif /* HAVE_STDARG_H */ - -#if !HAVE_VASPRINTF -#if HAVE_STDLIB_H -#include <stdlib.h> /* For malloc(3). */ -#endif /* HAVE_STDLIB_H */ -#ifdef VA_COPY -#undef VA_COPY -#endif /* defined(VA_COPY) */ -#ifdef VA_END_COPY -#undef VA_END_COPY -#endif /* defined(VA_END_COPY) */ -#if HAVE_VA_COPY -#define VA_COPY(dest, src) va_copy(dest, src) -#define VA_END_COPY(ap) va_end(ap) -#elif HAVE___VA_COPY -#define VA_COPY(dest, src) __va_copy(dest, src) -#define VA_END_COPY(ap) va_end(ap) -#else -#define VA_COPY(dest, src) (void)mymemcpy(&dest, &src, sizeof(va_list)) -#define VA_END_COPY(ap) /* No-op. */ -#define NEED_MYMEMCPY 1 -static void *mymemcpy(void *, void *, size_t); -#endif /* HAVE_VA_COPY */ -#endif /* !HAVE_VASPRINTF */ - -#if !HAVE_VSNPRINTF -#include <limits.h> /* For *_MAX. */ -#if HAVE_INTTYPES_H -#include <inttypes.h> /* For intmax_t (if not defined in <stdint.h>). */ -#endif /* HAVE_INTTYPES_H */ -#if HAVE_LOCALE_H -#include <locale.h> /* For localeconv(3). */ -#endif /* HAVE_LOCALE_H */ -#if HAVE_STDDEF_H -#include <stddef.h> /* For ptrdiff_t. */ -#endif /* HAVE_STDDEF_H */ -#if HAVE_STDINT_H -#include <stdint.h> /* For intmax_t. */ -#endif /* HAVE_STDINT_H */ - -/* Support for unsigned long long int. We may also need ULLONG_MAX. */ -#ifndef ULONG_MAX /* We may need ULONG_MAX as a fallback. */ -#ifdef UINT_MAX -#define ULONG_MAX UINT_MAX -#else -#define ULONG_MAX INT_MAX -#endif /* defined(UINT_MAX) */ -#endif /* !defined(ULONG_MAX) */ -#ifdef ULLONG -#undef ULLONG -#endif /* defined(ULLONG) */ -#if HAVE_UNSIGNED_LONG_LONG_INT -#define ULLONG unsigned long long int -#ifndef ULLONG_MAX -#define ULLONG_MAX ULONG_MAX -#endif /* !defined(ULLONG_MAX) */ -#else -#define ULLONG unsigned long int -#ifdef ULLONG_MAX -#undef ULLONG_MAX -#endif /* defined(ULLONG_MAX) */ -#define ULLONG_MAX ULONG_MAX -#endif /* HAVE_LONG_LONG_INT */ - -/* Support for uintmax_t. We also need UINTMAX_MAX. */ -#ifdef UINTMAX_T -#undef UINTMAX_T -#endif /* defined(UINTMAX_T) */ -#if HAVE_UINTMAX_T || defined(uintmax_t) -#define UINTMAX_T uintmax_t -#ifndef UINTMAX_MAX -#define UINTMAX_MAX ULLONG_MAX -#endif /* !defined(UINTMAX_MAX) */ -#else -#define UINTMAX_T ULLONG -#ifdef UINTMAX_MAX -#undef UINTMAX_MAX -#endif /* defined(UINTMAX_MAX) */ -#define UINTMAX_MAX ULLONG_MAX -#endif /* HAVE_UINTMAX_T || defined(uintmax_t) */ - -/* Support for long double. */ -#ifndef LDOUBLE -#if HAVE_LONG_DOUBLE -#define LDOUBLE long double -#else -#define LDOUBLE double -#endif /* HAVE_LONG_DOUBLE */ -#endif /* !defined(LDOUBLE) */ - -/* Support for long long int. */ -#ifndef LLONG -#if HAVE_LONG_LONG_INT -#define LLONG long long int -#else -#define LLONG long int -#endif /* HAVE_LONG_LONG_INT */ -#endif /* !defined(LLONG) */ - -/* Support for intmax_t. */ -#ifndef INTMAX_T -#if HAVE_INTMAX_T || defined(intmax_t) -#define INTMAX_T intmax_t -#else -#define INTMAX_T LLONG -#endif /* HAVE_INTMAX_T || defined(intmax_t) */ -#endif /* !defined(INTMAX_T) */ - -/* Support for uintptr_t. */ -#ifndef UINTPTR_T -#if HAVE_UINTPTR_T || defined(uintptr_t) -#define UINTPTR_T uintptr_t -#else -#define UINTPTR_T unsigned long int -#endif /* HAVE_UINTPTR_T || defined(uintptr_t) */ -#endif /* !defined(UINTPTR_T) */ - -/* Support for ptrdiff_t. */ -#ifndef PTRDIFF_T -#if HAVE_PTRDIFF_T || defined(ptrdiff_t) -#define PTRDIFF_T ptrdiff_t -#else -#define PTRDIFF_T long int -#endif /* HAVE_PTRDIFF_T || defined(ptrdiff_t) */ -#endif /* !defined(PTRDIFF_T) */ - -/* - * We need an unsigned integer type corresponding to ptrdiff_t (cf. C99: - * 7.19.6.1, 7). However, we'll simply use PTRDIFF_T and convert it to an - * unsigned type if necessary. This should work just fine in practice. - */ -#ifndef UPTRDIFF_T -#define UPTRDIFF_T PTRDIFF_T -#endif /* !defined(UPTRDIFF_T) */ - -/* - * We need a signed integer type corresponding to size_t (cf. C99: 7.19.6.1, 7). - * However, we'll simply use size_t and convert it to a signed type if - * necessary. This should work just fine in practice. - */ -#ifndef SSIZE_T -#define SSIZE_T size_t -#endif /* !defined(SSIZE_T) */ - -/* Either ERANGE or E2BIG should be available everywhere. */ -#ifndef ERANGE -#define ERANGE E2BIG -#endif /* !defined(ERANGE) */ -#ifndef EOVERFLOW -#define EOVERFLOW ERANGE -#endif /* !defined(EOVERFLOW) */ - -/* - * Buffer size to hold the octal string representation of UINT128_MAX without - * nul-termination ("3777777777777777777777777777777777777777777"). - */ -#ifdef MAX_CONVERT_LENGTH -#undef MAX_CONVERT_LENGTH -#endif /* defined(MAX_CONVERT_LENGTH) */ -#define MAX_CONVERT_LENGTH 43 - -/* Format read states. */ -#define PRINT_S_DEFAULT 0 -#define PRINT_S_FLAGS 1 -#define PRINT_S_WIDTH 2 -#define PRINT_S_DOT 3 -#define PRINT_S_PRECISION 4 -#define PRINT_S_MOD 5 -#define PRINT_S_CONV 6 - -/* Format flags. */ -#define PRINT_F_MINUS (1 << 0) -#define PRINT_F_PLUS (1 << 1) -#define PRINT_F_SPACE (1 << 2) -#define PRINT_F_NUM (1 << 3) -#define PRINT_F_ZERO (1 << 4) -#define PRINT_F_QUOTE (1 << 5) -#define PRINT_F_UP (1 << 6) -#define PRINT_F_UNSIGNED (1 << 7) -#define PRINT_F_TYPE_G (1 << 8) -#define PRINT_F_TYPE_E (1 << 9) - -/* Conversion flags. */ -#define PRINT_C_CHAR 1 -#define PRINT_C_SHORT 2 -#define PRINT_C_LONG 3 -#define PRINT_C_LLONG 4 -#define PRINT_C_LDOUBLE 5 -#define PRINT_C_SIZE 6 -#define PRINT_C_PTRDIFF 7 -#define PRINT_C_INTMAX 8 - -#ifndef MAX -#define MAX(x, y) ((x >= y) ? x : y) -#endif /* !defined(MAX) */ -#ifndef CHARTOINT -#define CHARTOINT(ch) (ch - '0') -#endif /* !defined(CHARTOINT) */ -#ifndef ISDIGIT -#define ISDIGIT(ch) ('0' <= (unsigned char)ch && (unsigned char)ch <= '9') -#endif /* !defined(ISDIGIT) */ -#ifndef ISNAN -#define ISNAN(x) (x != x) -#endif /* !defined(ISNAN) */ -#ifndef ISINF -#define ISINF(x) (x != 0.0 && x + x == x) -#endif /* !defined(ISINF) */ - -#ifdef OUTCHAR -#undef OUTCHAR -#endif /* defined(OUTCHAR) */ -#define OUTCHAR(str, len, size, ch) \ -do { \ - if (len + 1 < size) \ - str[len] = ch; \ - (len)++; \ -} while (/* CONSTCOND */ 0) - -static void fmtstr(char *, size_t *, size_t, const char *, int, int, int); -static void fmtint(char *, size_t *, size_t, INTMAX_T, int, int, int, int); -static void fmtflt(char *, size_t *, size_t, LDOUBLE, int, int, int, int *); -static void printsep(char *, size_t *, size_t); -static int getnumsep(int); -static int getexponent(LDOUBLE); -static int convert(UINTMAX_T, char *, size_t, int, int); -static UINTMAX_T cast(LDOUBLE); -static UINTMAX_T myround(LDOUBLE); -static LDOUBLE mypow10(int); - -int -util_vsnprintf(char *str, size_t size, const char *format, va_list args) -{ - LDOUBLE fvalue; - INTMAX_T value; - unsigned char cvalue; - const char *strvalue; - INTMAX_T *intmaxptr; - PTRDIFF_T *ptrdiffptr; - SSIZE_T *sizeptr; - LLONG *llongptr; - long int *longptr; - int *intptr; - short int *shortptr; - signed char *charptr; - size_t len = 0; - int overflow = 0; - int base = 0; - int cflags = 0; - int flags = 0; - int width = 0; - int precision = -1; - int state = PRINT_S_DEFAULT; - char ch = *format++; - - /* - * C99 says: "If `n' is zero, nothing is written, and `s' may be a null - * pointer." (7.19.6.5, 2) We're forgiving and allow a NULL pointer - * even if a size larger than zero was specified. At least NetBSD's - * snprintf(3) does the same, as well as other versions of this file. - * (Though some of these versions will write to a non-NULL buffer even - * if a size of zero was specified, which violates the standard.) - */ - if (str == NULL && size != 0) - size = 0; - - while (ch != '\0') - switch (state) { - case PRINT_S_DEFAULT: - if (ch == '%') - state = PRINT_S_FLAGS; - else - OUTCHAR(str, len, size, ch); - ch = *format++; - break; - case PRINT_S_FLAGS: - switch (ch) { - case '-': - flags |= PRINT_F_MINUS; - ch = *format++; - break; - case '+': - flags |= PRINT_F_PLUS; - ch = *format++; - break; - case ' ': - flags |= PRINT_F_SPACE; - ch = *format++; - break; - case '#': - flags |= PRINT_F_NUM; - ch = *format++; - break; - case '0': - flags |= PRINT_F_ZERO; - ch = *format++; - break; - case '\'': /* SUSv2 flag (not in C99). */ - flags |= PRINT_F_QUOTE; - ch = *format++; - break; - default: - state = PRINT_S_WIDTH; - break; - } - break; - case PRINT_S_WIDTH: - if (ISDIGIT(ch)) { - ch = CHARTOINT(ch); - if (width > (INT_MAX - ch) / 10) { - overflow = 1; - goto out; - } - width = 10 * width + ch; - ch = *format++; - } else if (ch == '*') { - /* - * C99 says: "A negative field width argument is - * taken as a `-' flag followed by a positive - * field width." (7.19.6.1, 5) - */ - if ((width = va_arg(args, int)) < 0) { - flags |= PRINT_F_MINUS; - width = -width; - } - ch = *format++; - state = PRINT_S_DOT; - } else - state = PRINT_S_DOT; - break; - case PRINT_S_DOT: - if (ch == '.') { - state = PRINT_S_PRECISION; - ch = *format++; - } else - state = PRINT_S_MOD; - break; - case PRINT_S_PRECISION: - if (precision == -1) - precision = 0; - if (ISDIGIT(ch)) { - ch = CHARTOINT(ch); - if (precision > (INT_MAX - ch) / 10) { - overflow = 1; - goto out; - } - precision = 10 * precision + ch; - ch = *format++; - } else if (ch == '*') { - /* - * C99 says: "A negative precision argument is - * taken as if the precision were omitted." - * (7.19.6.1, 5) - */ - if ((precision = va_arg(args, int)) < 0) - precision = -1; - ch = *format++; - state = PRINT_S_MOD; - } else - state = PRINT_S_MOD; - break; - case PRINT_S_MOD: - switch (ch) { - case 'h': - ch = *format++; - if (ch == 'h') { /* It's a char. */ - ch = *format++; - cflags = PRINT_C_CHAR; - } else - cflags = PRINT_C_SHORT; - break; - case 'l': - ch = *format++; - if (ch == 'l') { /* It's a long long. */ - ch = *format++; - cflags = PRINT_C_LLONG; - } else - cflags = PRINT_C_LONG; - break; - case 'L': - cflags = PRINT_C_LDOUBLE; - ch = *format++; - break; - case 'j': - cflags = PRINT_C_INTMAX; - ch = *format++; - break; - case 't': - cflags = PRINT_C_PTRDIFF; - ch = *format++; - break; - case 'z': - cflags = PRINT_C_SIZE; - ch = *format++; - break; - } - state = PRINT_S_CONV; - break; - case PRINT_S_CONV: - switch (ch) { - case 'd': - /* FALLTHROUGH */ - case 'i': - switch (cflags) { - case PRINT_C_CHAR: - value = (signed char)va_arg(args, int); - break; - case PRINT_C_SHORT: - value = (short int)va_arg(args, int); - break; - case PRINT_C_LONG: - value = va_arg(args, long int); - break; - case PRINT_C_LLONG: - value = va_arg(args, LLONG); - break; - case PRINT_C_SIZE: - value = va_arg(args, SSIZE_T); - break; - case PRINT_C_INTMAX: - value = va_arg(args, INTMAX_T); - break; - case PRINT_C_PTRDIFF: - value = va_arg(args, PTRDIFF_T); - break; - default: - value = va_arg(args, int); - break; - } - fmtint(str, &len, size, value, 10, width, - precision, flags); - break; - case 'X': - flags |= PRINT_F_UP; - /* FALLTHROUGH */ - case 'x': - base = 16; - /* FALLTHROUGH */ - case 'o': - if (base == 0) - base = 8; - /* FALLTHROUGH */ - case 'u': - if (base == 0) - base = 10; - flags |= PRINT_F_UNSIGNED; - switch (cflags) { - case PRINT_C_CHAR: - value = (unsigned char)va_arg(args, - unsigned int); - break; - case PRINT_C_SHORT: - value = (unsigned short int)va_arg(args, - unsigned int); - break; - case PRINT_C_LONG: - value = va_arg(args, unsigned long int); - break; - case PRINT_C_LLONG: - value = va_arg(args, ULLONG); - break; - case PRINT_C_SIZE: - value = va_arg(args, size_t); - break; - case PRINT_C_INTMAX: - value = va_arg(args, UINTMAX_T); - break; - case PRINT_C_PTRDIFF: - value = va_arg(args, UPTRDIFF_T); - break; - default: - value = va_arg(args, unsigned int); - break; - } - fmtint(str, &len, size, value, base, width, - precision, flags); - break; - case 'A': - /* Not yet supported, we'll use "%F". */ - /* FALLTHROUGH */ - case 'F': - flags |= PRINT_F_UP; - case 'a': - /* Not yet supported, we'll use "%f". */ - /* FALLTHROUGH */ - case 'f': - if (cflags == PRINT_C_LDOUBLE) - fvalue = va_arg(args, LDOUBLE); - else - fvalue = va_arg(args, double); - fmtflt(str, &len, size, fvalue, width, - precision, flags, &overflow); - if (overflow) - goto out; - break; - case 'E': - flags |= PRINT_F_UP; - /* FALLTHROUGH */ - case 'e': - flags |= PRINT_F_TYPE_E; - if (cflags == PRINT_C_LDOUBLE) - fvalue = va_arg(args, LDOUBLE); - else - fvalue = va_arg(args, double); - fmtflt(str, &len, size, fvalue, width, - precision, flags, &overflow); - if (overflow) - goto out; - break; - case 'G': - flags |= PRINT_F_UP; - /* FALLTHROUGH */ - case 'g': - flags |= PRINT_F_TYPE_G; - if (cflags == PRINT_C_LDOUBLE) - fvalue = va_arg(args, LDOUBLE); - else - fvalue = va_arg(args, double); - /* - * If the precision is zero, it is treated as - * one (cf. C99: 7.19.6.1, 8). - */ - if (precision == 0) - precision = 1; - fmtflt(str, &len, size, fvalue, width, - precision, flags, &overflow); - if (overflow) - goto out; - break; - case 'c': - cvalue = (unsigned char)va_arg(args, int); - OUTCHAR(str, len, size, cvalue); - break; - case 's': - strvalue = va_arg(args, char *); - fmtstr(str, &len, size, strvalue, width, - precision, flags); - break; - case 'p': - /* - * C99 says: "The value of the pointer is - * converted to a sequence of printing - * characters, in an implementation-defined - * manner." (C99: 7.19.6.1, 8) - */ - if ((strvalue = va_arg(args, void *)) == NULL) - /* - * We use the glibc format. BSD prints - * "0x0", SysV "0". - */ - fmtstr(str, &len, size, "(nil)", width, - -1, flags); - else { - /* - * We use the BSD/glibc format. SysV - * omits the "0x" prefix (which we emit - * using the PRINT_F_NUM flag). - */ - flags |= PRINT_F_NUM; - flags |= PRINT_F_UNSIGNED; - fmtint(str, &len, size, - (UINTPTR_T)strvalue, 16, width, - precision, flags); - } - break; - case 'n': - switch (cflags) { - case PRINT_C_CHAR: - charptr = va_arg(args, signed char *); - *charptr = (signed char)len; - break; - case PRINT_C_SHORT: - shortptr = va_arg(args, short int *); - *shortptr = (short int)len; - break; - case PRINT_C_LONG: - longptr = va_arg(args, long int *); - *longptr = (long int)len; - break; - case PRINT_C_LLONG: - llongptr = va_arg(args, LLONG *); - *llongptr = (LLONG)len; - break; - case PRINT_C_SIZE: - /* - * C99 says that with the "z" length - * modifier, "a following `n' conversion - * specifier applies to a pointer to a - * signed integer type corresponding to - * size_t argument." (7.19.6.1, 7) - */ - sizeptr = va_arg(args, SSIZE_T *); - *sizeptr = len; - break; - case PRINT_C_INTMAX: - intmaxptr = va_arg(args, INTMAX_T *); - *intmaxptr = len; - break; - case PRINT_C_PTRDIFF: - ptrdiffptr = va_arg(args, PTRDIFF_T *); - *ptrdiffptr = len; - break; - default: - intptr = va_arg(args, int *); - *intptr = (int)len; - break; - } - break; - case '%': /* Print a "%" character verbatim. */ - OUTCHAR(str, len, size, ch); - break; - default: /* Skip other characters. */ - break; - } - ch = *format++; - state = PRINT_S_DEFAULT; - base = cflags = flags = width = 0; - precision = -1; - break; - } -out: - if (len < size) - str[len] = '\0'; - else if (size > 0) - str[size - 1] = '\0'; - - if (overflow || len >= INT_MAX) { - return -1; - } - return (int)len; -} - -static void -fmtstr(char *str, size_t *len, size_t size, const char *value, int width, - int precision, int flags) -{ - int padlen, strln; /* Amount to pad. */ - int noprecision = (precision == -1); - - if (value == NULL) /* We're forgiving. */ - value = "(null)"; - - /* If a precision was specified, don't read the string past it. */ - for (strln = 0; value[strln] != '\0' && - (noprecision || strln < precision); strln++) - continue; - - if ((padlen = width - strln) < 0) - padlen = 0; - if (flags & PRINT_F_MINUS) /* Left justify. */ - padlen = -padlen; - - while (padlen > 0) { /* Leading spaces. */ - OUTCHAR(str, *len, size, ' '); - padlen--; - } - while (*value != '\0' && (noprecision || precision-- > 0)) { - OUTCHAR(str, *len, size, *value); - value++; - } - while (padlen < 0) { /* Trailing spaces. */ - OUTCHAR(str, *len, size, ' '); - padlen++; - } -} - -static void -fmtint(char *str, size_t *len, size_t size, INTMAX_T value, int base, int width, - int precision, int flags) -{ - UINTMAX_T uvalue; - char iconvert[MAX_CONVERT_LENGTH]; - char sign = 0; - char hexprefix = 0; - int spadlen = 0; /* Amount to space pad. */ - int zpadlen = 0; /* Amount to zero pad. */ - int pos; - int separators = (flags & PRINT_F_QUOTE); - int noprecision = (precision == -1); - - if (flags & PRINT_F_UNSIGNED) - uvalue = value; - else { - uvalue = (value >= 0) ? value : -value; - if (value < 0) - sign = '-'; - else if (flags & PRINT_F_PLUS) /* Do a sign. */ - sign = '+'; - else if (flags & PRINT_F_SPACE) - sign = ' '; - } - - pos = convert(uvalue, iconvert, sizeof(iconvert), base, - flags & PRINT_F_UP); - - if (flags & PRINT_F_NUM && uvalue != 0) { - /* - * C99 says: "The result is converted to an `alternative form'. - * For `o' conversion, it increases the precision, if and only - * if necessary, to force the first digit of the result to be a - * zero (if the value and precision are both 0, a single 0 is - * printed). For `x' (or `X') conversion, a nonzero result has - * `0x' (or `0X') prefixed to it." (7.19.6.1, 6) - */ - switch (base) { - case 8: - if (precision <= pos) - precision = pos + 1; - break; - case 16: - hexprefix = (flags & PRINT_F_UP) ? 'X' : 'x'; - break; - } - } - - if (separators) /* Get the number of group separators we'll print. */ - separators = getnumsep(pos); - - zpadlen = precision - pos - separators; - spadlen = width /* Minimum field width. */ - - separators /* Number of separators. */ - - MAX(precision, pos) /* Number of integer digits. */ - - ((sign != 0) ? 1 : 0) /* Will we print a sign? */ - - ((hexprefix != 0) ? 2 : 0); /* Will we print a prefix? */ - - if (zpadlen < 0) - zpadlen = 0; - if (spadlen < 0) - spadlen = 0; - - /* - * C99 says: "If the `0' and `-' flags both appear, the `0' flag is - * ignored. For `d', `i', `o', `u', `x', and `X' conversions, if a - * precision is specified, the `0' flag is ignored." (7.19.6.1, 6) - */ - if (flags & PRINT_F_MINUS) /* Left justify. */ - spadlen = -spadlen; - else if (flags & PRINT_F_ZERO && noprecision) { - zpadlen += spadlen; - spadlen = 0; - } - while (spadlen > 0) { /* Leading spaces. */ - OUTCHAR(str, *len, size, ' '); - spadlen--; - } - if (sign != 0) /* Sign. */ - OUTCHAR(str, *len, size, sign); - if (hexprefix != 0) { /* A "0x" or "0X" prefix. */ - OUTCHAR(str, *len, size, '0'); - OUTCHAR(str, *len, size, hexprefix); - } - while (zpadlen > 0) { /* Leading zeros. */ - OUTCHAR(str, *len, size, '0'); - zpadlen--; - } - while (pos > 0) { /* The actual digits. */ - pos--; - OUTCHAR(str, *len, size, iconvert[pos]); - if (separators > 0 && pos > 0 && pos % 3 == 0) - printsep(str, len, size); - } - while (spadlen < 0) { /* Trailing spaces. */ - OUTCHAR(str, *len, size, ' '); - spadlen++; - } -} - -static void -fmtflt(char *str, size_t *len, size_t size, LDOUBLE fvalue, int width, - int precision, int flags, int *overflow) -{ - LDOUBLE ufvalue; - UINTMAX_T intpart; - UINTMAX_T fracpart; - UINTMAX_T mask; - const char *infnan = NULL; - char iconvert[MAX_CONVERT_LENGTH]; - char fconvert[MAX_CONVERT_LENGTH]; - char econvert[4]; /* "e-12" (without nul-termination). */ - char esign = 0; - char sign = 0; - int leadfraczeros = 0; - int exponent = 0; - int emitpoint = 0; - int omitzeros = 0; - int omitcount = 0; - int padlen = 0; - int epos = 0; - int fpos = 0; - int ipos = 0; - int separators = (flags & PRINT_F_QUOTE); - int estyle = (flags & PRINT_F_TYPE_E); -#if HAVE_LOCALECONV && HAVE_LCONV_DECIMAL_POINT - struct lconv *lc = localeconv(); -#endif /* HAVE_LOCALECONV && HAVE_LCONV_DECIMAL_POINT */ - - /* - * AIX' man page says the default is 0, but C99 and at least Solaris' - * and NetBSD's man pages say the default is 6, and sprintf(3) on AIX - * defaults to 6. - */ - if (precision == -1) - precision = 6; - - if (fvalue < 0.0) - sign = '-'; - else if (flags & PRINT_F_PLUS) /* Do a sign. */ - sign = '+'; - else if (flags & PRINT_F_SPACE) - sign = ' '; - - if (ISNAN(fvalue)) - infnan = (flags & PRINT_F_UP) ? "NAN" : "nan"; - else if (ISINF(fvalue)) - infnan = (flags & PRINT_F_UP) ? "INF" : "inf"; - - if (infnan != NULL) { - if (sign != 0) - iconvert[ipos++] = sign; - while (*infnan != '\0') - iconvert[ipos++] = *infnan++; - fmtstr(str, len, size, iconvert, width, ipos, flags); - return; - } - - /* "%e" (or "%E") or "%g" (or "%G") conversion. */ - if (flags & PRINT_F_TYPE_E || flags & PRINT_F_TYPE_G) { - if (flags & PRINT_F_TYPE_G) { - /* - * For "%g" (and "%G") conversions, the precision - * specifies the number of significant digits, which - * includes the digits in the integer part. The - * conversion will or will not be using "e-style" (like - * "%e" or "%E" conversions) depending on the precision - * and on the exponent. However, the exponent can be - * affected by rounding the converted value, so we'll - * leave this decision for later. Until then, we'll - * assume that we're going to do an "e-style" conversion - * (in order to get the exponent calculated). For - * "e-style", the precision must be decremented by one. - */ - precision--; - /* - * For "%g" (and "%G") conversions, trailing zeros are - * removed from the fractional portion of the result - * unless the "#" flag was specified. - */ - if (!(flags & PRINT_F_NUM)) - omitzeros = 1; - } - exponent = getexponent(fvalue); - estyle = 1; - } - -again: - /* - * Sorry, we only support 9, 19, or 38 digits (that is, the number of - * digits of the 32-bit, the 64-bit, or the 128-bit UINTMAX_MAX value - * minus one) past the decimal point due to our conversion method. - */ - switch (sizeof(UINTMAX_T)) { - case 16: - if (precision > 38) - precision = 38; - break; - case 8: - if (precision > 19) - precision = 19; - break; - default: - if (precision > 9) - precision = 9; - break; - } - - ufvalue = (fvalue >= 0.0) ? fvalue : -fvalue; - if (estyle) /* We want exactly one integer digit. */ - ufvalue /= mypow10(exponent); - - if ((intpart = cast(ufvalue)) == UINTMAX_MAX) { - *overflow = 1; - return; - } - - /* - * Factor of ten with the number of digits needed for the fractional - * part. For example, if the precision is 3, the mask will be 1000. - */ - mask = (UINTMAX_T)mypow10(precision); - /* - * We "cheat" by converting the fractional part to integer by - * multiplying by a factor of ten. - */ - if ((fracpart = myround(mask * (ufvalue - intpart))) >= mask) { - /* - * For example, ufvalue = 2.99962, intpart = 2, and mask = 1000 - * (because precision = 3). Now, myround(1000 * 0.99962) will - * return 1000. So, the integer part must be incremented by one - * and the fractional part must be set to zero. - */ - intpart++; - fracpart = 0; - if (estyle && intpart == 10) { - /* - * The value was rounded up to ten, but we only want one - * integer digit if using "e-style". So, the integer - * part must be set to one and the exponent must be - * incremented by one. - */ - intpart = 1; - exponent++; - } - } - - /* - * Now that we know the real exponent, we can check whether or not to - * use "e-style" for "%g" (and "%G") conversions. If we don't need - * "e-style", the precision must be adjusted and the integer and - * fractional parts must be recalculated from the original value. - * - * C99 says: "Let P equal the precision if nonzero, 6 if the precision - * is omitted, or 1 if the precision is zero. Then, if a conversion - * with style `E' would have an exponent of X: - * - * - if P > X >= -4, the conversion is with style `f' (or `F') and - * precision P - (X + 1). - * - * - otherwise, the conversion is with style `e' (or `E') and precision - * P - 1." (7.19.6.1, 8) - * - * Note that we had decremented the precision by one. - */ - if (flags & PRINT_F_TYPE_G && estyle && - precision + 1 > exponent && exponent >= -4) { - precision -= exponent; - estyle = 0; - goto again; - } - - if (estyle) { - if (exponent < 0) { - exponent = -exponent; - esign = '-'; - } else - esign = '+'; - - /* - * Convert the exponent. The sizeof(econvert) is 4. So, the - * econvert buffer can hold e.g. "e+99" and "e-99". We don't - * support an exponent which contains more than two digits. - * Therefore, the following stores are safe. - */ - epos = convert(exponent, econvert, 2, 10, 0); - /* - * C99 says: "The exponent always contains at least two digits, - * and only as many more digits as necessary to represent the - * exponent." (7.19.6.1, 8) - */ - if (epos == 1) - econvert[epos++] = '0'; - econvert[epos++] = esign; - econvert[epos++] = (flags & PRINT_F_UP) ? 'E' : 'e'; - } - - /* Convert the integer part and the fractional part. */ - ipos = convert(intpart, iconvert, sizeof(iconvert), 10, 0); - if (fracpart != 0) /* convert() would return 1 if fracpart == 0. */ - fpos = convert(fracpart, fconvert, sizeof(fconvert), 10, 0); - - leadfraczeros = precision - fpos; - - if (omitzeros) { - if (fpos > 0) /* Omit trailing fractional part zeros. */ - while (omitcount < fpos && fconvert[omitcount] == '0') - omitcount++; - else { /* The fractional part is zero, omit it completely. */ - omitcount = precision; - leadfraczeros = 0; - } - precision -= omitcount; - } - - /* - * Print a decimal point if either the fractional part is non-zero - * and/or the "#" flag was specified. - */ - if (precision > 0 || flags & PRINT_F_NUM) - emitpoint = 1; - if (separators) /* Get the number of group separators we'll print. */ - separators = getnumsep(ipos); - - padlen = width /* Minimum field width. */ - - ipos /* Number of integer digits. */ - - epos /* Number of exponent characters. */ - - precision /* Number of fractional digits. */ - - separators /* Number of group separators. */ - - (emitpoint ? 1 : 0) /* Will we print a decimal point? */ - - ((sign != 0) ? 1 : 0); /* Will we print a sign character? */ - - if (padlen < 0) - padlen = 0; - - /* - * C99 says: "If the `0' and `-' flags both appear, the `0' flag is - * ignored." (7.19.6.1, 6) - */ - if (flags & PRINT_F_MINUS) /* Left justifty. */ - padlen = -padlen; - else if (flags & PRINT_F_ZERO && padlen > 0) { - if (sign != 0) { /* Sign. */ - OUTCHAR(str, *len, size, sign); - sign = 0; - } - while (padlen > 0) { /* Leading zeros. */ - OUTCHAR(str, *len, size, '0'); - padlen--; - } - } - while (padlen > 0) { /* Leading spaces. */ - OUTCHAR(str, *len, size, ' '); - padlen--; - } - if (sign != 0) /* Sign. */ - OUTCHAR(str, *len, size, sign); - while (ipos > 0) { /* Integer part. */ - ipos--; - OUTCHAR(str, *len, size, iconvert[ipos]); - if (separators > 0 && ipos > 0 && ipos % 3 == 0) - printsep(str, len, size); - } - if (emitpoint) { /* Decimal point. */ -#if HAVE_LOCALECONV && HAVE_LCONV_DECIMAL_POINT - if (lc->decimal_point != NULL && *lc->decimal_point != '\0') - OUTCHAR(str, *len, size, *lc->decimal_point); - else /* We'll always print some decimal point character. */ -#endif /* HAVE_LOCALECONV && HAVE_LCONV_DECIMAL_POINT */ - OUTCHAR(str, *len, size, '.'); - } - while (leadfraczeros > 0) { /* Leading fractional part zeros. */ - OUTCHAR(str, *len, size, '0'); - leadfraczeros--; - } - while (fpos > omitcount) { /* The remaining fractional part. */ - fpos--; - OUTCHAR(str, *len, size, fconvert[fpos]); - } - while (epos > 0) { /* Exponent. */ - epos--; - OUTCHAR(str, *len, size, econvert[epos]); - } - while (padlen < 0) { /* Trailing spaces. */ - OUTCHAR(str, *len, size, ' '); - padlen++; - } -} - -static void -printsep(char *str, size_t *len, size_t size) -{ -#if HAVE_LOCALECONV && HAVE_LCONV_THOUSANDS_SEP - struct lconv *lc = localeconv(); - int i; - - if (lc->thousands_sep != NULL) - for (i = 0; lc->thousands_sep[i] != '\0'; i++) - OUTCHAR(str, *len, size, lc->thousands_sep[i]); - else -#endif /* HAVE_LOCALECONV && HAVE_LCONV_THOUSANDS_SEP */ - OUTCHAR(str, *len, size, ','); -} - -static int -getnumsep(int digits) -{ - int separators = (digits - ((digits % 3 == 0) ? 1 : 0)) / 3; -#if HAVE_LOCALECONV && HAVE_LCONV_THOUSANDS_SEP - int strln; - struct lconv *lc = localeconv(); - - /* We support an arbitrary separator length (including zero). */ - if (lc->thousands_sep != NULL) { - for (strln = 0; lc->thousands_sep[strln] != '\0'; strln++) - continue; - separators *= strln; - } -#endif /* HAVE_LOCALECONV && HAVE_LCONV_THOUSANDS_SEP */ - return separators; -} - -static int -getexponent(LDOUBLE value) -{ - LDOUBLE tmp = (value >= 0.0) ? value : -value; - int exponent = 0; - - /* - * We check for 99 > exponent > -99 in order to work around possible - * endless loops which could happen (at least) in the second loop (at - * least) if we're called with an infinite value. However, we checked - * for infinity before calling this function using our ISINF() macro, so - * this might be somewhat paranoid. - */ - while (tmp < 1.0 && tmp > 0.0 && --exponent > -99) - tmp *= 10; - while (tmp >= 10.0 && ++exponent < 99) - tmp /= 10; - - return exponent; -} - -static int -convert(UINTMAX_T value, char *buf, size_t size, int base, int caps) -{ - const char *digits = caps ? "0123456789ABCDEF" : "0123456789abcdef"; - size_t pos = 0; - - /* We return an unterminated buffer with the digits in reverse order. */ - do { - buf[pos++] = digits[value % base]; - value /= base; - } while (value != 0 && pos < size); - - return (int)pos; -} - -static UINTMAX_T -cast(LDOUBLE value) -{ - UINTMAX_T result; - - /* - * We check for ">=" and not for ">" because if UINTMAX_MAX cannot be - * represented exactly as an LDOUBLE value (but is less than LDBL_MAX), - * it may be increased to the nearest higher representable value for the - * comparison (cf. C99: 6.3.1.4, 2). It might then equal the LDOUBLE - * value although converting the latter to UINTMAX_T would overflow. - */ - if (value >= UINTMAX_MAX) - return UINTMAX_MAX; - - result = (UINTMAX_T)value; - /* - * At least on NetBSD/sparc64 3.0.2 and 4.99.30, casting long double to - * an integer type converts e.g. 1.9 to 2 instead of 1 (which violates - * the standard). Sigh. - */ - return (result <= value) ? result : result - 1; -} - -static UINTMAX_T -myround(LDOUBLE value) -{ - UINTMAX_T intpart = cast(value); - - return ((value -= intpart) < 0.5) ? intpart : intpart + 1; -} - -static LDOUBLE -mypow10(int exponent) -{ - LDOUBLE result = 1; - - while (exponent > 0) { - result *= 10; - exponent--; - } - while (exponent < 0) { - result /= 10; - exponent++; - } - return result; -} -#endif /* !HAVE_VSNPRINTF */ - -#if !HAVE_VASPRINTF -#if NEED_MYMEMCPY -void * -mymemcpy(void *dst, void *src, size_t len) -{ - const char *from = src; - char *to = dst; - - /* No need for optimization, we use this only to replace va_copy(3). */ - while (len-- > 0) - *to++ = *from++; - return dst; -} -#endif /* NEED_MYMEMCPY */ - -int -util_vasprintf(char **ret, const char *format, va_list ap) -{ - size_t size; - int len; - va_list aq; - - VA_COPY(aq, ap); - len = vsnprintf(NULL, 0, format, aq); - VA_END_COPY(aq); - if (len < 0 || (*ret = malloc(size = len + 1)) == NULL) - return -1; - return vsnprintf(*ret, size, format, ap); -} -#endif /* !HAVE_VASPRINTF */ - -#if !HAVE_SNPRINTF -#if HAVE_STDARG_H -int -util_snprintf(char *str, size_t size, const char *format, ...) -#else -int -util_snprintf(va_alist) va_dcl -#endif /* HAVE_STDARG_H */ -{ -#if !HAVE_STDARG_H - char *str; - size_t size; - char *format; -#endif /* HAVE_STDARG_H */ - va_list ap; - int len; - - VA_START(ap, format); - VA_SHIFT(ap, str, char *); - VA_SHIFT(ap, size, size_t); - VA_SHIFT(ap, format, const char *); - len = vsnprintf(str, size, format, ap); - va_end(ap); - return len; -} -#endif /* !HAVE_SNPRINTF */ - -#if !HAVE_ASPRINTF -#if HAVE_STDARG_H -int -util_asprintf(char **ret, const char *format, ...) -#else -int -util_asprintf(va_alist) va_dcl -#endif /* HAVE_STDARG_H */ -{ -#if !HAVE_STDARG_H - char **ret; - char *format; -#endif /* HAVE_STDARG_H */ - va_list ap; - int len; - - VA_START(ap, format); - VA_SHIFT(ap, ret, char **); - VA_SHIFT(ap, format, const char *); - len = vasprintf(ret, format, ap); - va_end(ap); - return len; -} -#endif /* !HAVE_ASPRINTF */ -#else /* Dummy declaration to avoid empty translation unit warnings. */ -int main(void); -#endif /* !HAVE_SNPRINTF || !HAVE_VSNPRINTF || !HAVE_ASPRINTF || [...] */ - - -/* vim: set joinspaces textwidth=80: */ diff --git a/src/gallium/auxiliary/util/u_split_prim.h b/src/gallium/auxiliary/util/u_split_prim.h index 7f80fc12700..5afb7d9a920 100644 --- a/src/gallium/auxiliary/util/u_split_prim.h +++ b/src/gallium/auxiliary/util/u_split_prim.h @@ -23,7 +23,7 @@ struct util_split_prim { uint edgeflag_off:1; }; -static INLINE void +static inline void util_split_prim_init(struct util_split_prim *s, unsigned mode, unsigned start, unsigned count) { @@ -41,7 +41,7 @@ util_split_prim_init(struct util_split_prim *s, s->repeat_first = 0; } -static INLINE boolean +static inline boolean util_split_prim_next(struct util_split_prim *s, unsigned max_verts) { int repeat = 0; diff --git a/src/gallium/auxiliary/util/u_sse.h b/src/gallium/auxiliary/util/u_sse.h index d4f51912a2d..7f8e5a1a3cf 100644 --- a/src/gallium/auxiliary/util/u_sse.h +++ b/src/gallium/auxiliary/util/u_sse.h @@ -51,7 +51,7 @@ union m128i { uint ui[4]; }; -static INLINE void u_print_epi8(const char *name, __m128i r) +static inline void u_print_epi8(const char *name, __m128i r) { union { __m128i m; ubyte ub[16]; } u; u.m = r; @@ -80,7 +80,7 @@ static INLINE void u_print_epi8(const char *name, __m128i r) u.ub[12], u.ub[13], u.ub[14], u.ub[15]); } -static INLINE void u_print_epi16(const char *name, __m128i r) +static inline void u_print_epi16(const char *name, __m128i r) { union { __m128i m; ushort us[8]; } u; u.m = r; @@ -99,7 +99,7 @@ static INLINE void u_print_epi16(const char *name, __m128i r) u.us[4], u.us[5], u.us[6], u.us[7]); } -static INLINE void u_print_epi32(const char *name, __m128i r) +static inline void u_print_epi32(const char *name, __m128i r) { union { __m128i m; uint ui[4]; } u; u.m = r; @@ -113,7 +113,7 @@ static INLINE void u_print_epi32(const char *name, __m128i r) u.ui[0], u.ui[1], u.ui[2], u.ui[3]); } -static INLINE void u_print_ps(const char *name, __m128 r) +static inline void u_print_ps(const char *name, __m128 r) { union { __m128 m; float f[4]; } u; u.m = r; @@ -179,7 +179,7 @@ _mm_shuffle_epi8(__m128i a, __m128i mask) * _mm_mullo_epi32() intrinsic as to not justify adding an sse4 * dependency at this point. */ -static INLINE __m128i mm_mullo_epi32(const __m128i a, const __m128i b) +static inline __m128i mm_mullo_epi32(const __m128i a, const __m128i b) { __m128i a4 = _mm_srli_epi64(a, 32); /* shift by one dword */ __m128i b4 = _mm_srli_epi64(b, 32); /* shift by one dword */ @@ -204,7 +204,7 @@ static INLINE __m128i mm_mullo_epi32(const __m128i a, const __m128i b) } -static INLINE void +static inline void transpose4_epi32(const __m128i * restrict a, const __m128i * restrict b, const __m128i * restrict c, diff --git a/src/gallium/auxiliary/util/u_string.h b/src/gallium/auxiliary/util/u_string.h index dc89c4400bc..f7ab09c8f1c 100644 --- a/src/gallium/auxiliary/util/u_string.h +++ b/src/gallium/auxiliary/util/u_string.h @@ -35,13 +35,14 @@ #ifndef U_STRING_H_ #define U_STRING_H_ -#if !defined(_MSC_VER) && !defined(XF86_LIBC_H) +#if !defined(XF86_LIBC_H) #include <stdio.h> #endif #include <stddef.h> #include <stdarg.h> #include "pipe/p_compiler.h" +#include "util/macros.h" // PRINTFLIKE #ifdef __cplusplus @@ -54,7 +55,7 @@ extern "C" { #else -static INLINE char * +static inline char * util_strchrnul(const char *s, char c) { for (; *s && *s != c; ++s); @@ -64,18 +65,44 @@ util_strchrnul(const char *s, char c) #endif -#ifdef _MSC_VER +#ifdef _WIN32 -int util_vsnprintf(char *, size_t, const char *, va_list); -int util_snprintf(char *str, size_t size, const char *format, ...); +static inline int +util_vsnprintf(char *str, size_t size, const char *format, va_list ap) +{ + /* We need to use _vscprintf to calculate the length as vsnprintf returns -1 + * if the number of characters to write is greater than count. + */ + va_list ap_copy; + int ret; + va_copy(ap_copy, ap); + ret = _vsnprintf(str, size, format, ap); + if (ret < 0) { + ret = _vscprintf(format, ap_copy); + } + return ret; +} + +static inline int + PRINTFLIKE(3, 4) +util_snprintf(char *str, size_t size, const char *format, ...) +{ + va_list ap; + int ret; + va_start(ap, format); + ret = util_vsnprintf(str, size, format, ap); + va_end(ap); + return ret; +} -static INLINE void +static inline void util_vsprintf(char *str, const char *format, va_list ap) { util_vsnprintf(str, (size_t)-1, format, ap); } -static INLINE void +static inline void + PRINTFLIKE(2, 3) util_sprintf(char *str, const char *format, ...) { va_list ap; @@ -84,7 +111,7 @@ util_sprintf(char *str, const char *format, ...) va_end(ap); } -static INLINE char * +static inline char * util_strchr(const char *s, char c) { char *p = util_strchrnul(s, c); @@ -92,7 +119,7 @@ util_strchr(const char *s, char c) return *p ? p : NULL; } -static INLINE char* +static inline char* util_strncat(char *dst, const char *src, size_t n) { char *p = dst + strlen(dst); @@ -106,7 +133,7 @@ util_strncat(char *dst, const char *src, size_t n) return dst; } -static INLINE int +static inline int util_strcmp(const char *s1, const char *s2) { unsigned char u1, u2; @@ -122,7 +149,7 @@ util_strcmp(const char *s1, const char *s2) return 0; } -static INLINE int +static inline int util_strncmp(const char *s1, const char *s2, size_t n) { unsigned char u1, u2; @@ -138,7 +165,7 @@ util_strncmp(const char *s1, const char *s2, size_t n) return 0; } -static INLINE char * +static inline char * util_strstr(const char *haystack, const char *needle) { const char *p = haystack; @@ -152,7 +179,7 @@ util_strstr(const char *haystack, const char *needle) return NULL; } -static INLINE void * +static inline void * util_memmove(void *dest, const void *src, size_t n) { char *p = (char *)dest; @@ -199,7 +226,7 @@ struct util_strbuf }; -static INLINE void +static inline void util_strbuf_init(struct util_strbuf *sbuf, char *str, size_t size) { sbuf->str = str; @@ -209,7 +236,7 @@ util_strbuf_init(struct util_strbuf *sbuf, char *str, size_t size) } -static INLINE void +static inline void util_strbuf_printf(struct util_strbuf *sbuf, const char *format, ...) { if(sbuf->left > 1) { diff --git a/src/gallium/auxiliary/util/u_surfaces.h b/src/gallium/auxiliary/util/u_surfaces.h index 1605215cb88..b84694c540b 100644 --- a/src/gallium/auxiliary/util/u_surfaces.h +++ b/src/gallium/auxiliary/util/u_surfaces.h @@ -50,7 +50,7 @@ util_surfaces_do_get(struct util_surfaces *us, unsigned surface_struct_size, struct pipe_surface **res); /* fast inline path for the very common case */ -static INLINE boolean +static inline boolean util_surfaces_get(struct util_surfaces *us, unsigned surface_struct_size, struct pipe_context *ctx, struct pipe_resource *pt, unsigned level, unsigned layer, @@ -70,7 +70,7 @@ util_surfaces_get(struct util_surfaces *us, unsigned surface_struct_size, return util_surfaces_do_get(us, surface_struct_size, ctx, pt, level, layer, res); } -static INLINE struct pipe_surface * +static inline struct pipe_surface * util_surfaces_peek(struct util_surfaces *us, struct pipe_resource *pt, unsigned level, unsigned layer) { if(!us->u.pv) @@ -84,7 +84,7 @@ util_surfaces_peek(struct util_surfaces *us, struct pipe_resource *pt, unsigned void util_surfaces_do_detach(struct util_surfaces *us, struct pipe_surface *ps); -static INLINE void +static inline void util_surfaces_detach(struct util_surfaces *us, struct pipe_surface *ps) { if(likely(ps->texture->target == PIPE_TEXTURE_2D || ps->texture->target == PIPE_TEXTURE_RECT)) diff --git a/src/gallium/auxiliary/util/u_tile.h b/src/gallium/auxiliary/util/u_tile.h index a33d7f7722b..dc1f568a8e5 100644 --- a/src/gallium/auxiliary/util/u_tile.h +++ b/src/gallium/auxiliary/util/u_tile.h @@ -42,7 +42,7 @@ struct pipe_transfer; * * \return TRUE if tile is totally clipped, FALSE otherwise */ -static INLINE boolean +static inline boolean u_clip_tile(uint x, uint y, uint *w, uint *h, const struct pipe_box *box) { if ((int) x >= box->width) diff --git a/src/gallium/auxiliary/util/u_time.h b/src/gallium/auxiliary/util/u_time.h index 2bee1e00014..a5017d6bce2 100644 --- a/src/gallium/auxiliary/util/u_time.h +++ b/src/gallium/auxiliary/util/u_time.h @@ -60,7 +60,7 @@ struct util_time PIPE_DEPRECATED -static INLINE void +static inline void util_time_get(struct util_time *t) { t->counter = os_time_get(); @@ -71,7 +71,7 @@ util_time_get(struct util_time *t) * Return t2 = t1 + usecs */ PIPE_DEPRECATED -static INLINE void +static inline void util_time_add(const struct util_time *t1, int64_t usecs, struct util_time *t2) @@ -84,7 +84,7 @@ util_time_add(const struct util_time *t1, * Return difference between times, in microseconds */ PIPE_DEPRECATED -static INLINE int64_t +static inline int64_t util_time_diff(const struct util_time *t1, const struct util_time *t2) { @@ -98,7 +98,7 @@ util_time_diff(const struct util_time *t1, * Not publicly available because it does not take in account wrap-arounds. * Use util_time_timeout instead. */ -static INLINE int +static inline int _util_time_compare(const struct util_time *t1, const struct util_time *t2) { @@ -115,7 +115,7 @@ _util_time_compare(const struct util_time *t1, * Returns non-zero when the timeout expires. */ PIPE_DEPRECATED -static INLINE boolean +static inline boolean util_time_timeout(const struct util_time *start, const struct util_time *end, const struct util_time *curr) @@ -128,7 +128,7 @@ util_time_timeout(const struct util_time *start, * Return current time in microseconds */ PIPE_DEPRECATED -static INLINE int64_t +static inline int64_t util_time_micros(void) { return os_time_get(); @@ -136,7 +136,7 @@ util_time_micros(void) PIPE_DEPRECATED -static INLINE void +static inline void util_time_sleep(int64_t usecs) { os_time_sleep(usecs); diff --git a/src/gallium/auxiliary/util/u_transfer.c b/src/gallium/auxiliary/util/u_transfer.c index 71da35d6d39..4cb524d5cb1 100644 --- a/src/gallium/auxiliary/util/u_transfer.c +++ b/src/gallium/auxiliary/util/u_transfer.c @@ -90,7 +90,7 @@ void u_default_transfer_unmap( struct pipe_context *pipe, } -static INLINE struct u_resource * +static inline struct u_resource * u_resource( struct pipe_resource *res ) { return (struct u_resource *)res; diff --git a/src/gallium/auxiliary/util/u_video.h b/src/gallium/auxiliary/util/u_video.h index b4743d13fbf..ddc00216105 100644 --- a/src/gallium/auxiliary/util/u_video.h +++ b/src/gallium/auxiliary/util/u_video.h @@ -40,7 +40,7 @@ extern "C" { #include "util/u_debug.h" #include "util/u_math.h" -static INLINE enum pipe_video_format +static inline enum pipe_video_format u_reduce_video_profile(enum pipe_video_profile profile) { switch (profile) @@ -68,12 +68,19 @@ u_reduce_video_profile(enum pipe_video_profile profile) case PIPE_VIDEO_PROFILE_MPEG4_AVC_HIGH444: return PIPE_VIDEO_FORMAT_MPEG4_AVC; + case PIPE_VIDEO_PROFILE_HEVC_MAIN: + case PIPE_VIDEO_PROFILE_HEVC_MAIN_10: + case PIPE_VIDEO_PROFILE_HEVC_MAIN_STILL: + case PIPE_VIDEO_PROFILE_HEVC_MAIN_12: + case PIPE_VIDEO_PROFILE_HEVC_MAIN_444: + return PIPE_VIDEO_FORMAT_HEVC; + default: return PIPE_VIDEO_FORMAT_UNKNOWN; } } -static INLINE void +static inline void u_copy_nv12_to_yv12(void *const *destination_data, uint32_t const *destination_pitches, int src_plane, int src_field, @@ -99,7 +106,7 @@ u_copy_nv12_to_yv12(void *const *destination_data, } } -static INLINE void +static inline void u_copy_yv12_to_nv12(void *const *destination_data, uint32_t const *destination_pitches, int src_plane, int src_field, @@ -122,7 +129,7 @@ u_copy_yv12_to_nv12(void *const *destination_data, } } -static INLINE void +static inline void u_copy_swap422_packed(void *const *destination_data, uint32_t const *destination_pitches, int src_plane, int src_field, @@ -147,7 +154,7 @@ u_copy_swap422_packed(void *const *destination_data, } } -static INLINE uint32_t +static inline uint32_t u_get_h264_level(uint32_t width, uint32_t height, uint32_t *max_reference) { uint32_t max_dpb_mbs; diff --git a/src/gallium/auxiliary/vl/vl_compositor.c b/src/gallium/auxiliary/vl/vl_compositor.c index 69839e61386..afe53063b48 100644 --- a/src/gallium/auxiliary/vl/vl_compositor.c +++ b/src/gallium/auxiliary/vl/vl_compositor.c @@ -538,7 +538,7 @@ cleanup_buffers(struct vl_compositor *c) pipe_resource_reference(&c->vertex_buf.buffer, NULL); } -static INLINE struct u_rect +static inline struct u_rect default_rect(struct vl_compositor_layer *layer) { struct pipe_resource *res = layer->sampler_views[0]->texture; @@ -546,21 +546,21 @@ default_rect(struct vl_compositor_layer *layer) return rect; } -static INLINE struct vertex2f +static inline struct vertex2f calc_topleft(struct vertex2f size, struct u_rect rect) { struct vertex2f res = { rect.x0 / size.x, rect.y0 / size.y }; return res; } -static INLINE struct vertex2f +static inline struct vertex2f calc_bottomright(struct vertex2f size, struct u_rect rect) { struct vertex2f res = { rect.x1 / size.x, rect.y1 / size.y }; return res; } -static INLINE void +static inline void calc_src_and_dst(struct vl_compositor_layer *layer, unsigned width, unsigned height, struct u_rect src, struct u_rect dst) { @@ -658,7 +658,7 @@ gen_rect_verts(struct vertex2f *vb, struct vl_compositor_layer *layer) vb[19].y = layer->colors[3].w; } -static INLINE struct u_rect +static inline struct u_rect calc_drawn_area(struct vl_compositor_state *s, struct vl_compositor_layer *layer) { struct vertex2f tl, br; diff --git a/src/gallium/auxiliary/vl/vl_mpeg12_bitstream.c b/src/gallium/auxiliary/vl/vl_mpeg12_bitstream.c index abb3780f61e..52ce6c416aa 100644 --- a/src/gallium/auxiliary/vl/vl_mpeg12_bitstream.c +++ b/src/gallium/auxiliary/vl/vl_mpeg12_bitstream.c @@ -533,7 +533,7 @@ static struct dct_coeff tbl_B14_DC[1 << 17]; static struct dct_coeff tbl_B14_AC[1 << 17]; static struct dct_coeff tbl_B15[1 << 17]; -static INLINE void +static inline void init_dct_coeff_table(struct dct_coeff *dst, const struct dct_coeff_compressed *src, unsigned size, bool is_DC) { @@ -594,7 +594,7 @@ init_dct_coeff_table(struct dct_coeff *dst, const struct dct_coeff_compressed *s } } -static INLINE void +static inline void init_tables() { vl_vlc_init_table(tbl_B1, Elements(tbl_B1), macroblock_address_increment, Elements(macroblock_address_increment)); @@ -611,19 +611,19 @@ init_tables() init_dct_coeff_table(tbl_B15, dct_coeff_tbl_one, Elements(dct_coeff_tbl_one), false); } -static INLINE int +static inline int DIV2DOWN(int todiv) { return (todiv&~1)/2; } -static INLINE int +static inline int DIV2UP(int todiv) { return (todiv+1)/2; } -static INLINE void +static inline void motion_vector(struct vl_mpg12_bs *bs, int r, int s, int dmv, short delta[2], short dmvector[2]) { int t; @@ -647,7 +647,7 @@ motion_vector(struct vl_mpg12_bs *bs, int r, int s, int dmv, short delta[2], sho } } -static INLINE int +static inline int wrap(short f, int shift) { if (f < (-16 << shift)) @@ -658,7 +658,7 @@ wrap(short f, int shift) return f; } -static INLINE void +static inline void motion_vector_frame(struct vl_mpg12_bs *bs, int s, struct pipe_mpeg12_macroblock *mb) { int dmv = mb->macroblock_modes.bits.frame_motion_type == PIPE_MPEG12_MO_TYPE_DUAL_PRIME; @@ -682,7 +682,7 @@ motion_vector_frame(struct vl_mpg12_bs *bs, int s, struct pipe_mpeg12_macroblock } } -static INLINE void +static inline void motion_vector_field(struct vl_mpg12_bs *bs, int s, struct pipe_mpeg12_macroblock *mb) { int dmv = mb->macroblock_modes.bits.field_motion_type == PIPE_MPEG12_MO_TYPE_DUAL_PRIME; @@ -701,12 +701,12 @@ motion_vector_field(struct vl_mpg12_bs *bs, int s, struct pipe_mpeg12_macroblock } } -static INLINE void +static inline void reset_predictor(struct vl_mpg12_bs *bs) { bs->pred_dc[0] = bs->pred_dc[1] = bs->pred_dc[2] = 0; } -static INLINE void +static inline void decode_dct(struct vl_mpg12_bs *bs, struct pipe_mpeg12_macroblock *mb, int scale) { static const unsigned blk2cc[] = { 0, 0, 0, 0, 1, 2 }; @@ -805,7 +805,7 @@ entry: vl_vlc_eatbits(&bs->vlc, 1); } -static INLINE void +static inline void decode_slice(struct vl_mpg12_bs *bs, struct pipe_video_buffer *target) { struct pipe_mpeg12_macroblock mb; @@ -929,6 +929,7 @@ decode_slice(struct vl_mpg12_bs *bs, struct pipe_video_buffer *target) mb.PMV[1][0][0] = mb.PMV[0][0][0]; mb.PMV[1][0][1] = mb.PMV[0][0][1]; assert(extra); + (void) extra; } else if (mb.macroblock_type & PIPE_MPEG12_MB_TYPE_INTRA || !(mb.macroblock_type & (PIPE_MPEG12_MB_TYPE_MOTION_FORWARD | PIPE_MPEG12_MB_TYPE_MOTION_BACKWARD))) { diff --git a/src/gallium/auxiliary/vl/vl_mpeg12_decoder.c b/src/gallium/auxiliary/vl/vl_mpeg12_decoder.c index 8579460e070..b7009837293 100644 --- a/src/gallium/auxiliary/vl/vl_mpeg12_decoder.c +++ b/src/gallium/auxiliary/vl/vl_mpeg12_decoder.c @@ -84,6 +84,9 @@ static const unsigned const_empty_block_mask_420[3][2][2] = { struct video_buffer_private { + struct list_head list; + struct pipe_video_buffer *video_buffer; + struct pipe_sampler_view *sampler_view_planes[VL_NUM_COMPONENTS]; struct pipe_surface *surfaces[VL_MAX_SURFACES]; @@ -99,6 +102,8 @@ destroy_video_buffer_private(void *private) struct video_buffer_private *priv = private; unsigned i; + list_del(&priv->list); + for (i = 0; i < VL_NUM_COMPONENTS; ++i) pipe_sampler_view_reference(&priv->sampler_view_planes[i], NULL); @@ -126,6 +131,9 @@ get_video_buffer_private(struct vl_mpeg12_decoder *dec, struct pipe_video_buffer priv = CALLOC_STRUCT(video_buffer_private); + list_add(&priv->list, &dec->buffer_privates); + priv->video_buffer = buf; + sv = buf->get_sampler_view_planes(buf); for (i = 0; i < VL_NUM_COMPONENTS; ++i) if (sv[i]) @@ -141,6 +149,18 @@ get_video_buffer_private(struct vl_mpeg12_decoder *dec, struct pipe_video_buffer return priv; } +static void +free_video_buffer_privates(struct vl_mpeg12_decoder *dec) +{ + struct video_buffer_private *priv, *next; + + LIST_FOR_EACH_ENTRY_SAFE(priv, next, &dec->buffer_privates, list) { + struct pipe_video_buffer *buf = priv->video_buffer; + + vl_video_buffer_set_associated_data(buf, &dec->base, NULL, NULL); + } +} + static bool init_zscan_buffer(struct vl_mpeg12_decoder *dec, struct vl_mpeg12_buffer *buffer) { @@ -297,7 +317,7 @@ cleanup_mc_buffer(struct vl_mpeg12_buffer *buf) vl_mc_cleanup_buffer(&buf->mc[i]); } -static INLINE void +static inline void MacroBlockTypeToPipeWeights(const struct pipe_mpeg12_macroblock *mb, unsigned weights[2]) { assert(mb); @@ -332,7 +352,7 @@ MacroBlockTypeToPipeWeights(const struct pipe_mpeg12_macroblock *mb, unsigned we } } -static INLINE struct vl_motionvector +static inline struct vl_motionvector MotionVectorToPipe(const struct pipe_mpeg12_macroblock *mb, unsigned vector, unsigned field_select_mask, unsigned weight) { @@ -383,7 +403,7 @@ MotionVectorToPipe(const struct pipe_mpeg12_macroblock *mb, unsigned vector, return mv; } -static INLINE void +static inline void UploadYcbcrBlocks(struct vl_mpeg12_decoder *dec, struct vl_mpeg12_buffer *buf, const struct pipe_mpeg12_macroblock *mb) @@ -464,6 +484,8 @@ vl_mpeg12_destroy(struct pipe_video_codec *decoder) assert(decoder); + free_video_buffer_privates(dec); + /* Asserted in softpipe_delete_fs_state() for some reason */ dec->context->bind_vs_state(dec->context, NULL); dec->context->bind_fs_state(dec->context, NULL); @@ -1187,6 +1209,8 @@ vl_create_mpeg12_decoder(struct pipe_context *context, if (!init_pipe_state(dec)) goto error_pipe_state; + list_inithead(&dec->buffer_privates); + return &dec->base; error_pipe_state: diff --git a/src/gallium/auxiliary/vl/vl_mpeg12_decoder.h b/src/gallium/auxiliary/vl/vl_mpeg12_decoder.h index 2a604054387..505dd675f66 100644 --- a/src/gallium/auxiliary/vl/vl_mpeg12_decoder.h +++ b/src/gallium/auxiliary/vl/vl_mpeg12_decoder.h @@ -30,6 +30,8 @@ #include "pipe/p_video_codec.h" +#include "util/list.h" + #include "vl_mpeg12_bitstream.h" #include "vl_zscan.h" #include "vl_idct.h" @@ -77,6 +79,8 @@ struct vl_mpeg12_decoder unsigned current_buffer; struct vl_mpeg12_buffer *dec_buffers[4]; + + struct list_head buffer_privates; }; struct vl_mpeg12_buffer diff --git a/src/gallium/auxiliary/vl/vl_rbsp.h b/src/gallium/auxiliary/vl/vl_rbsp.h index 2e3da8e1d28..7867238c49e 100644 --- a/src/gallium/auxiliary/vl/vl_rbsp.h +++ b/src/gallium/auxiliary/vl/vl_rbsp.h @@ -48,7 +48,7 @@ struct vl_rbsp { /** * Initialize the RBSP object */ -static INLINE void vl_rbsp_init(struct vl_rbsp *rbsp, struct vl_vlc *nal, unsigned num_bits) +static inline void vl_rbsp_init(struct vl_rbsp *rbsp, struct vl_vlc *nal, unsigned num_bits) { unsigned bits_left = vl_vlc_bits_left(nal); @@ -71,7 +71,7 @@ static INLINE void vl_rbsp_init(struct vl_rbsp *rbsp, struct vl_vlc *nal, unsign /** * Make at least 16 more bits available */ -static INLINE void vl_rbsp_fillbits(struct vl_rbsp *rbsp) +static inline void vl_rbsp_fillbits(struct vl_rbsp *rbsp) { unsigned valid = vl_vlc_valid_bits(&rbsp->nal); unsigned i, bits; @@ -108,7 +108,7 @@ static INLINE void vl_rbsp_fillbits(struct vl_rbsp *rbsp) /** * Return an unsigned integer from the first n bits */ -static INLINE unsigned vl_rbsp_u(struct vl_rbsp *rbsp, unsigned n) +static inline unsigned vl_rbsp_u(struct vl_rbsp *rbsp, unsigned n) { if (n == 0) return 0; @@ -120,7 +120,7 @@ static INLINE unsigned vl_rbsp_u(struct vl_rbsp *rbsp, unsigned n) /** * Return an unsigned exponential Golomb encoded integer */ -static INLINE unsigned vl_rbsp_ue(struct vl_rbsp *rbsp) +static inline unsigned vl_rbsp_ue(struct vl_rbsp *rbsp) { unsigned bits = 0; @@ -134,7 +134,7 @@ static INLINE unsigned vl_rbsp_ue(struct vl_rbsp *rbsp) /** * Return an signed exponential Golomb encoded integer */ -static INLINE signed vl_rbsp_se(struct vl_rbsp *rbsp) +static inline signed vl_rbsp_se(struct vl_rbsp *rbsp) { signed codeNum = vl_rbsp_ue(rbsp); if (codeNum & 1) @@ -146,7 +146,7 @@ static INLINE signed vl_rbsp_se(struct vl_rbsp *rbsp) /** * Are more data available in the RBSP ? */ -static INLINE bool vl_rbsp_more_data(struct vl_rbsp *rbsp) +static inline bool vl_rbsp_more_data(struct vl_rbsp *rbsp) { unsigned bits, value; diff --git a/src/gallium/auxiliary/vl/vl_vlc.h b/src/gallium/auxiliary/vl/vl_vlc.h index 2f905956dbf..7821b8be0a1 100644 --- a/src/gallium/auxiliary/vl/vl_vlc.h +++ b/src/gallium/auxiliary/vl/vl_vlc.h @@ -65,7 +65,7 @@ struct vl_vlc_compressed /** * initalize and decompress a lookup table */ -static INLINE void +static inline void vl_vlc_init_table(struct vl_vlc_entry *dst, unsigned dst_size, const struct vl_vlc_compressed *src, unsigned src_size) { unsigned i, bits = util_logbase2(dst_size); @@ -87,7 +87,7 @@ vl_vlc_init_table(struct vl_vlc_entry *dst, unsigned dst_size, const struct vl_v /** * switch over to next input buffer */ -static INLINE void +static inline void vl_vlc_next_input(struct vl_vlc *vlc) { unsigned len = vlc->sizes[0]; @@ -112,7 +112,7 @@ vl_vlc_next_input(struct vl_vlc *vlc) /** * align the data pointer to the next dword */ -static INLINE void +static inline void vl_vlc_align_data_ptr(struct vl_vlc *vlc) { /* align the data pointer */ @@ -126,7 +126,7 @@ vl_vlc_align_data_ptr(struct vl_vlc *vlc) /** * fill the bit buffer, so that at least 32 bits are valid */ -static INLINE void +static inline void vl_vlc_fillbits(struct vl_vlc *vlc) { assert(vlc); @@ -175,7 +175,7 @@ vl_vlc_fillbits(struct vl_vlc *vlc) /** * initialize vlc structure and start reading from first input buffer */ -static INLINE void +static inline void vl_vlc_init(struct vl_vlc *vlc, unsigned num_inputs, const void *const *inputs, const unsigned *sizes) { @@ -203,7 +203,7 @@ vl_vlc_init(struct vl_vlc *vlc, unsigned num_inputs, /** * number of bits still valid in bit buffer */ -static INLINE unsigned +static inline unsigned vl_vlc_valid_bits(struct vl_vlc *vlc) { return 32 - vlc->invalid_bits; @@ -212,7 +212,7 @@ vl_vlc_valid_bits(struct vl_vlc *vlc) /** * number of bits left over all inbut buffers */ -static INLINE unsigned +static inline unsigned vl_vlc_bits_left(struct vl_vlc *vlc) { signed bytes_left = vlc->end - vlc->data; @@ -223,7 +223,7 @@ vl_vlc_bits_left(struct vl_vlc *vlc) /** * get num_bits from bit buffer without removing them */ -static INLINE unsigned +static inline unsigned vl_vlc_peekbits(struct vl_vlc *vlc, unsigned num_bits) { assert(vl_vlc_valid_bits(vlc) >= num_bits || vlc->data >= vlc->end); @@ -233,7 +233,7 @@ vl_vlc_peekbits(struct vl_vlc *vlc, unsigned num_bits) /** * remove num_bits from bit buffer */ -static INLINE void +static inline void vl_vlc_eatbits(struct vl_vlc *vlc, unsigned num_bits) { assert(vl_vlc_valid_bits(vlc) >= num_bits); @@ -245,7 +245,7 @@ vl_vlc_eatbits(struct vl_vlc *vlc, unsigned num_bits) /** * get num_bits from bit buffer with removing them */ -static INLINE unsigned +static inline unsigned vl_vlc_get_uimsbf(struct vl_vlc *vlc, unsigned num_bits) { unsigned value; @@ -261,7 +261,7 @@ vl_vlc_get_uimsbf(struct vl_vlc *vlc, unsigned num_bits) /** * treat num_bits as signed value and remove them from bit buffer */ -static INLINE signed +static inline signed vl_vlc_get_simsbf(struct vl_vlc *vlc, unsigned num_bits) { signed value; @@ -277,7 +277,7 @@ vl_vlc_get_simsbf(struct vl_vlc *vlc, unsigned num_bits) /** * lookup a value and length in a decompressed table */ -static INLINE int8_t +static inline int8_t vl_vlc_get_vlclbf(struct vl_vlc *vlc, const struct vl_vlc_entry *tbl, unsigned num_bits) { tbl += vl_vlc_peekbits(vlc, num_bits); @@ -288,7 +288,7 @@ vl_vlc_get_vlclbf(struct vl_vlc *vlc, const struct vl_vlc_entry *tbl, unsigned n /** * fast forward search for a specific byte value */ -static INLINE boolean +static inline boolean vl_vlc_search_byte(struct vl_vlc *vlc, unsigned num_bits, uint8_t value) { /* make sure we are on a byte boundary */ @@ -345,7 +345,7 @@ vl_vlc_search_byte(struct vl_vlc *vlc, unsigned num_bits, uint8_t value) /** * remove num_bits bits starting at pos from the bitbuffer */ -static INLINE void +static inline void vl_vlc_removebits(struct vl_vlc *vlc, unsigned pos, unsigned num_bits) { uint64_t lo = (vlc->buffer & (~0UL >> (pos + num_bits))) << num_bits; @@ -357,7 +357,7 @@ vl_vlc_removebits(struct vl_vlc *vlc, unsigned pos, unsigned num_bits) /** * limit the number of bits left for fetching */ -static INLINE void +static inline void vl_vlc_limit(struct vl_vlc *vlc, unsigned bits_left) { assert(bits_left <= vl_vlc_bits_left(vlc)); diff --git a/src/gallium/auxiliary/vl/vl_winsys_dri.c b/src/gallium/auxiliary/vl/vl_winsys_dri.c index 7e61b88e6b5..3b1b87f9523 100644 --- a/src/gallium/auxiliary/vl/vl_winsys_dri.c +++ b/src/gallium/auxiliary/vl/vl_winsys_dri.c @@ -37,6 +37,8 @@ #include <xf86drm.h> #include <errno.h> +#include "loader.h" + #include "pipe/p_screen.h" #include "pipe/p_context.h" #include "pipe/p_state.h" @@ -293,6 +295,16 @@ vl_screen_get_private(struct vl_screen *vscreen) return vscreen; } +static xcb_screen_t * +get_xcb_screen(xcb_screen_iterator_t iter, int screen) +{ + for (; iter.rem; --screen, xcb_screen_next(&iter)) + if (screen == 0) + return iter.data; + + return NULL; +} + struct vl_screen* vl_screen_create(Display *display, int screen) { @@ -334,8 +346,7 @@ vl_screen_create(Display *display, int screen) goto free_query; s = xcb_setup_roots_iterator(xcb_get_setup(scrn->conn)); - while (screen--) - xcb_screen_next(&s); + driverType = XCB_DRI2_DRIVER_TYPE_DRI; #ifdef DRI2DriverPrimeShift { @@ -351,7 +362,7 @@ vl_screen_create(Display *display, int screen) } #endif - connect_cookie = xcb_dri2_connect_unchecked(scrn->conn, s.data->root, driverType); + connect_cookie = xcb_dri2_connect_unchecked(scrn->conn, get_xcb_screen(s, screen)->root, driverType); connect = xcb_dri2_connect_reply(scrn->conn, connect_cookie, NULL); if (connect == NULL || connect->driver_name_length + connect->device_name_length == 0) goto free_connect; @@ -361,7 +372,7 @@ vl_screen_create(Display *display, int screen) if (!device_name) goto free_connect; memcpy(device_name, xcb_dri2_connect_device_name(connect), device_name_length); - fd = open(device_name, O_RDWR); + fd = loader_open_device(device_name); free(device_name); if (fd < 0) @@ -370,7 +381,7 @@ vl_screen_create(Display *display, int screen) if (drmGetMagic(fd, &magic)) goto free_connect; - authenticate_cookie = xcb_dri2_authenticate_unchecked(scrn->conn, s.data->root, magic); + authenticate_cookie = xcb_dri2_authenticate_unchecked(scrn->conn, get_xcb_screen(s, screen)->root, magic); authenticate = xcb_dri2_authenticate_reply(scrn->conn, authenticate_cookie, NULL); if (authenticate == NULL || !authenticate->authenticated) @@ -379,7 +390,7 @@ vl_screen_create(Display *display, int screen) #if GALLIUM_STATIC_TARGETS scrn->base.pscreen = dd_create_screen(fd); #else - if (pipe_loader_drm_probe_fd(&scrn->base.dev, fd, false)) + if (pipe_loader_drm_probe_fd(&scrn->base.dev, fd)) scrn->base.pscreen = pipe_loader_create_screen(scrn->base.dev, PIPE_SEARCH_DIR); #endif // GALLIUM_STATIC_TARGETS |