struct brw_context;
struct brw_inst;
struct brw_vs_prog_key;
-struct brw_vec4_prog_key;
+struct brw_vue_prog_key;
struct brw_wm_prog_key;
struct brw_wm_prog_data;
BRW_STATE_SF_VP,
BRW_STATE_CLIP_VP,
BRW_STATE_SAMPLER_STATE_TABLE,
+ BRW_STATE_VS_ATTRIB_WORKAROUNDS,
BRW_NUM_STATE_BITS
};
#define BRW_NEW_SF_VP (1ull << BRW_STATE_SF_VP)
#define BRW_NEW_CLIP_VP (1ull << BRW_STATE_CLIP_VP)
#define BRW_NEW_SAMPLER_STATE_TABLE (1ull << BRW_STATE_SAMPLER_STATE_TABLE)
+#define BRW_NEW_VS_ATTRIB_WORKAROUNDS (1ull << BRW_STATE_VS_ATTRIB_WORKAROUNDS)
struct brw_state_flags {
/** State update flags signalled by mesa internals */
* State update flags signalled as the result of brw_tracked_state updates
*/
uint64_t brw;
- /**
- * State update flags that used to be signalled by brw_state_cache.c
- * searches.
- *
- * Now almost all of that state is just streamed out on demand, but the
- * flags for those state blobs updating have stayed in the same bitfield.
- * brw_state_cache.c still flags BRW_NEW_*_PROG_DATA.
- */
- GLuint cache;
};
/** Subclass of Mesa vertex program */
uint32_t gather_texture_start;
uint32_t ubo_start;
uint32_t abo_start;
+ uint32_t image_start;
uint32_t shader_time_start;
/** @} */
} binding_table;
*/
unsigned dispatch_grf_start_reg;
+ bool use_alt_mode; /**< Use ALT floating point mode? Otherwise, IEEE. */
+
/* Pointers to tracked values (only valid once
* _mesa_load_state_parameters has been called at runtime).
*
/** @} */
} binding_table;
+ uint8_t computed_depth_mode;
+
bool no_8;
bool dual_src_blend;
bool uses_pos_offset;
};
-/* Note: brw_vec4_prog_data_compare() must be updated when adding fields to
+/* Note: brw_vue_prog_data_compare() must be updated when adding fields to
* this struct!
*/
-struct brw_vec4_prog_data {
+struct brw_vue_prog_data {
struct brw_stage_prog_data base;
struct brw_vue_map vue_map;
* is the size of the URB entry used for output.
*/
GLuint urb_entry_size;
+
+ bool simd8;
};
* struct!
*/
struct brw_vs_prog_data {
- struct brw_vec4_prog_data base;
+ struct brw_vue_prog_data base;
GLbitfield64 inputs_read;
/** Max number of atomic counter buffer objects in a shader */
#define BRW_MAX_ABO 16
+/** Max number of image uniforms in a shader */
+#define BRW_MAX_IMAGES 32
+
/**
* Max number of binding table entries used for stream output.
*
BRW_MAX_TEX_UNIT * 2 + /* normal, gather */ \
12 + /* ubo */ \
BRW_MAX_ABO + \
+ BRW_MAX_IMAGES + \
2 /* shader time, pull constants */)
#define SURF_INDEX_GEN6_SOL_BINDING(t) (t)
*/
struct brw_gs_prog_data
{
- struct brw_vec4_prog_data base;
+ struct brw_vue_prog_data base;
/**
* Size of an output vertex, measured in HWORDS (32 bytes).
struct brw_cache_item {
/**
* Effectively part of the key, cache_id identifies what kind of state
- * buffer is involved, and also which brw->state.dirty.cache flag should
- * be set when this cache item is chosen.
+ * buffer is involved, and also which dirty flag should set.
*/
enum brw_cache_id cache_id;
/** 32-bit hash of the key data */
/** Last index in bo with query data for this object. */
int last_index;
+
+ /** True if we know the batch has been flushed since we ended the query. */
+ bool flushed;
};
struct intel_sync_object {
drm_intel_bo *last_bo;
/** BO for post-sync nonzero writes for gen6 workaround. */
drm_intel_bo *workaround_bo;
- bool need_workaround_flush;
uint16_t emit, total;
uint16_t used, reserved_space;
enum brw_gpu_ring ring;
bool needs_sol_reset;
+ uint8_t pipe_controls_since_last_cs_stall;
+
struct {
uint16_t used;
int reloc_count;
struct gl_renderbuffer *rb,
bool layered,
unsigned unit);
- void (*update_null_renderbuffer_surface)(struct brw_context *brw,
- unsigned unit);
- void (*create_raw_surface)(struct brw_context *brw,
- drm_intel_bo *bo,
- uint32_t offset,
- uint32_t size,
- uint32_t *out_offset,
- bool rw);
void (*emit_buffer_surface_state)(struct brw_context *brw,
uint32_t *out_offset,
drm_intel_bo *bo,
unsigned surface_format,
unsigned buffer_size,
unsigned pitch,
- unsigned mocs,
bool rw);
+ void (*emit_null_surface_state)(struct brw_context *brw,
+ unsigned width,
+ unsigned height,
+ unsigned samples,
+ uint32_t *out_offset);
/**
* Send the appropriate state packets to configure depth, stencil, and
bool front_buffer_dirty;
/** Framerate throttling: @{ */
- drm_intel_bo *first_post_swapbuffers_batch;
- bool need_throttle;
+ drm_intel_bo *throttle_batch[2];
+
+ /* Limit the number of outstanding SwapBuffers by waiting for an earlier
+ * frame of rendering to complete. This gives a very precise cap to the
+ * latency between input and output such that rendering never gets more
+ * than a frame behind the user. (With the caveat that we technically are
+ * not using the SwapBuffers itself as a barrier but the first batch
+ * submitted afterwards, which may be immediately prior to the next
+ * SwapBuffers.)
+ */
+ bool need_swap_throttle;
+
+ /** General throttling, not caught by throttling between SwapBuffers */
+ bool need_flush_throttle;
/** @} */
GLuint stats_wm;
int gen;
int gt;
+ /* GT revision. This will be -1 if the revision couldn't be determined (eg,
+ * if the kernel doesn't support the query).
+ */
+ int revision;
bool is_g4x;
bool is_baytrail;
bool has_pln;
bool no_simd8;
bool use_rep_send;
+ bool scalar_vs;
/**
* Some versions of Gen hardware don't do centroid interpolation correctly
uint32_t pma_stall_bits;
struct {
- /** Does the current draw use the index buffer? */
- bool indexed;
-
- int start_vertex_location;
- int base_vertex_location;
+ /** The value of gl_BaseVertex for the current _mesa_prim. */
+ int gl_basevertex;
/**
* Buffer and offset used for GL_ARB_shader_draw_parameters
* the same VB packed over and over again.
*/
unsigned int start_vertex_bias;
+
+ /**
+ * Certain vertex attribute formats aren't natively handled by the
+ * hardware and require special VS code to fix up their values.
+ *
+ * These bitfields indicate which workarounds are needed.
+ */
+ uint8_t attrib_wa_flags[VERT_ATTRIB_MAX];
} vb;
struct {
* for each pipeline stage.
*/
int max_vs_threads;
+ int max_hs_threads;
+ int max_ds_threads;
int max_gs_threads;
int max_wm_threads;
GLuint min_vs_entries; /* Minimum number of VS entries */
GLuint max_vs_entries; /* Maximum number of VS entries */
+ GLuint max_hs_entries; /* Maximum number of HS entries */
+ GLuint max_ds_entries; /* Maximum number of DS entries */
GLuint max_gs_entries; /* Maximum number of GS entries */
GLuint nr_vs_entries;
/**
* Buffer object used in place of multisampled null render targets on
- * Gen6. See brw_update_null_renderbuffer_surface().
+ * Gen6. See brw_emit_null_surface_state().
*/
drm_intel_bo *multisampled_null_render_target_bo;
uint32_t fast_clear_op;
} perfmon;
int num_atoms;
- const struct brw_tracked_state **atoms;
+ const struct brw_tracked_state atoms[57];
/* If (INTEL_DEBUG & DEBUG_BATCH) */
struct {
struct intel_mipmap_tree *dst);
void brw_meta_fbo_stencil_blit(struct brw_context *brw,
+ struct gl_framebuffer *read_fb,
+ struct gl_framebuffer *draw_fb,
GLfloat srcX0, GLfloat srcY0,
GLfloat srcX1, GLfloat srcY1,
GLfloat dstX0, GLfloat dstY0,
/* brw_draw_upload.c */
unsigned brw_get_vertex_surface_type(struct brw_context *brw,
const struct gl_client_array *glarray);
-unsigned brw_get_index_type(GLenum type);
+
+static inline unsigned
+brw_get_index_type(GLenum type)
+{
+ assert((type == GL_UNSIGNED_BYTE)
+ || (type == GL_UNSIGNED_SHORT)
+ || (type == GL_UNSIGNED_INT));
+
+ /* The possible values for type are GL_UNSIGNED_BYTE (0x1401),
+ * GL_UNSIGNED_SHORT (0x1403), and GL_UNSIGNED_INT (0x1405) which we want
+ * to map to scale factors of 0, 1, and 2, respectively. These scale
+ * factors are then left-shfited by 8 to be in the correct position in the
+ * CMD_INDEX_BUFFER packet.
+ *
+ * Subtracting 0x1401 gives 0, 2, and 4. Shifting left by 7 afterwards
+ * gives 0x00000000, 0x00000100, and 0x00000200. These just happen to be
+ * the values the need to be written in the CMD_INDEX_BUFFER packet.
+ */
+ return (type - 0x1401) << 7;
+}
+
void brw_prepare_vertices(struct brw_context *brw);
/* brw_wm_surface_state.c */
void brw_upload_ubo_surfaces(struct brw_context *brw,
struct gl_shader *shader,
struct brw_stage_state *stage_state,
- struct brw_stage_prog_data *prog_data);
+ struct brw_stage_prog_data *prog_data,
+ bool dword_pitch);
void brw_upload_abo_surfaces(struct brw_context *brw,
struct gl_shader_program *prog,
struct brw_stage_state *stage_state,
/* brw_blorp_blit.cpp */
GLbitfield
brw_blorp_framebuffer(struct brw_context *brw,
+ struct gl_framebuffer *readFb,
+ struct gl_framebuffer *drawFb,
GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1,
GLbitfield mask, GLenum filter);
uint32_t get_hw_prim_for_gl_prim(int mode);
void
-brw_setup_vec4_key_clip_info(struct brw_context *brw,
- struct brw_vec4_prog_key *key,
- bool program_uses_clip_distance);
+brw_setup_vue_key_clip_info(struct brw_context *brw,
+ struct brw_vue_prog_key *key,
+ bool program_uses_clip_distance);
void
gen6_upload_push_constants(struct brw_context *brw,
struct brw_stage_state *stage_state,
enum aub_state_struct_type type);
-/* ================================================================
- * From linux kernel i386 header files, copes with odd sizes better
- * than COPY_DWORDS would:
- * XXX Put this in src/mesa/main/imports.h ???
- */
-#if defined(i386) || defined(__i386__)
-static inline void * __memcpy(void * to, const void * from, size_t n)
-{
- int d0, d1, d2;
- __asm__ __volatile__(
- "rep ; movsl\n\t"
- "testb $2,%b4\n\t"
- "je 1f\n\t"
- "movsw\n"
- "1:\ttestb $1,%b4\n\t"
- "je 2f\n\t"
- "movsb\n"
- "2:"
- : "=&c" (d0), "=&D" (d1), "=&S" (d2)
- :"0" (n/4), "q" (n),"1" ((long) to),"2" ((long) from)
- : "memory");
- return (to);
-}
-#else
-#define __memcpy(a,b,c) memcpy(a,b,c)
-#endif
-
#ifdef __cplusplus
}
#endif