Add vulkan driver for BDW

16 files changed, 221 insertions, 25 deletions

diff --git a/src/mesa/drivers/dri/common/dri_test.c b/src/mesa/drivers/dri/common/dri_test.c
index 57bfa5b9394..310e7617e2f 100644
--- a/src/mesa/drivers/dri/common/dri_test.c
+++ b/src/mesa/drivers/dri/common/dri_test.c
@@ -1,3 +1,4 @@
+#include <stdlib.h>
 #include "main/glheader.h"
 #include "main/compiler.h"
 #include "glapi/glapi.h"
@@ -33,12 +34,14 @@ _glapi_check_multithread(void)
 
 PUBLIC void
 _glapi_set_context(void *context)
-{}
+{
+        _glapi_Context = context;
+}
 
 PUBLIC void *
 _glapi_get_context(void)
 {
-	return 0;
+        return _glapi_Context;
 }
 
 PUBLIC void
@@ -84,7 +87,7 @@ _glapi_set_nop_handler(_glapi_nop_handler_proc func)
 PUBLIC struct _glapi_table *
 _glapi_new_nop_table(unsigned num_entries)
 {
-	return NULL;
+        return malloc(16);
 }
 
 #ifndef NO_MAIN
diff --git a/src/mesa/drivers/dri/i965/Makefile.am b/src/mesa/drivers/dri/i965/Makefile.am
index 235bcfeae02..cf2424e34b4 100644
--- a/src/mesa/drivers/dri/i965/Makefile.am
+++ b/src/mesa/drivers/dri/i965/Makefile.am
@@ -41,7 +41,7 @@ AM_CFLAGS = \
 
 AM_CXXFLAGS = $(AM_CFLAGS)
 
-noinst_LTLIBRARIES = libi965_dri.la
+noinst_LTLIBRARIES = libi965_dri.la libi965_compiler.la
 libi965_dri_la_SOURCES = $(i965_FILES)
 libi965_dri_la_LIBADD = $(INTEL_LIBS)
 
@@ -54,6 +54,15 @@ TEST_LIBS = \
         $(CLOCK_LIB) \
 	../common/libdri_test_stubs.la
 
+libi965_compiler_la_SOURCES = $(i965_FILES)
+libi965_compiler_la_LIBADD = $(INTEL_LIBS) \
+	../common/libdricommon.la \
+	../common/libmegadriver_stub.la \
+        ../../../libmesa.la \
+	$(DRI_LIB_DEPS) \
+        $(CLOCK_LIB) \
+	../common/libdri_test_stubs.la -lm
+
 TESTS = \
 	test_fs_cmod_propagation \
 	test_fs_saturate_propagation \
diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index fd7420a6c6f..18a30a5925e 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -830,6 +830,7 @@ brwCreateContext(gl_api api,
 
    intel_batchbuffer_init(brw);
 
+#if 0
    if (brw->gen >= 6) {
       /* Create a new hardware context.  Using a hardware context means that
        * our GPU state will be saved/restored on context switch, allowing us
@@ -848,6 +849,7 @@ brwCreateContext(gl_api api,
    }
 
    brw_init_state(brw);
+#endif
 
    intelInitExtensions(ctx);
 
@@ -909,8 +911,10 @@ brwCreateContext(gl_api api,
 
    _mesa_compute_version(ctx);
 
+#if 0
    _mesa_initialize_dispatch_tables(ctx);
    _mesa_initialize_vbo_vtxfmt(ctx);
+#endif
 
    if (ctx->Extensions.AMD_performance_monitor) {
       brw_init_performance_monitors(brw);
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 834aaa45737..2d4a7eab20b 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -359,6 +359,9 @@ struct brw_stage_prog_data {
       /** @} */
    } binding_table;
 
+   uint32_t *map_entries;
+   uint32_t *bind_map[4];
+
    GLuint nr_params;       /**< number of float params/constants */
    GLuint nr_pull_params;
 
@@ -1976,6 +1979,12 @@ gen6_upload_push_constants(struct brw_context *brw,
                            struct brw_stage_state *stage_state,
                            enum aub_state_struct_type type);
 
+struct intel_screen *intel_screen_create(int fd);
+void intel_screen_destroy(struct intel_screen *screen);
+
+struct brw_context *intel_context_create(struct intel_screen *screen);
+void intel_context_destroy(struct brw_context *brw);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h
index 83d7a3535e4..3c704ee9d08 100644
--- a/src/mesa/drivers/dri/i965/brw_defines.h
+++ b/src/mesa/drivers/dri/i965/brw_defines.h
@@ -55,6 +55,7 @@
 # define GEN7_3DPRIM_VERTEXBUFFER_ACCESS_SEQUENTIAL (0 << 8)
 # define GEN7_3DPRIM_VERTEXBUFFER_ACCESS_RANDOM     (1 << 8)
 
+#ifndef _3DPRIM_POINTLIST /* FIXME: Avoid clashing with defines from bdw_pack.h */
 #define _3DPRIM_POINTLIST         0x01
 #define _3DPRIM_LINELIST          0x02
 #define _3DPRIM_LINESTRIP         0x03
@@ -76,6 +77,7 @@
 #define _3DPRIM_LINESTRIP_BF      0x13
 #define _3DPRIM_LINESTRIP_CONT_BF 0x14
 #define _3DPRIM_TRIFAN_NOSTIPPLE  0x15
+#endif
 
 /* We use this offset to be able to pass native primitive types in struct
  * _mesa_prim::mode.  Native primitive types are BRW_PRIM_OFFSET +
diff --git a/src/mesa/drivers/dri/i965/brw_device_info.c b/src/mesa/drivers/dri/i965/brw_device_info.c
index 97243a47293..a07b86e60e2 100644
--- a/src/mesa/drivers/dri/i965/brw_device_info.c
+++ b/src/mesa/drivers/dri/i965/brw_device_info.c
@@ -353,3 +353,15 @@ brw_get_device_info(int devid, int revision)
 
    return devinfo;
 }
+
+const char *
+brw_get_device_name(int devid)
+{
+   switch (devid) {
+#undef CHIPSET
+#define CHIPSET(id, family, name) case id: return name;
+#include "pci_ids/i965_pci_ids.h"
+   default:
+      return NULL;
+   }
+}
diff --git a/src/mesa/drivers/dri/i965/brw_device_info.h b/src/mesa/drivers/dri/i965/brw_device_info.h
index 65c024ceeed..9192235fb0e 100644
--- a/src/mesa/drivers/dri/i965/brw_device_info.h
+++ b/src/mesa/drivers/dri/i965/brw_device_info.h
@@ -85,3 +85,4 @@ struct brw_device_info
 };
 
 const struct brw_device_info *brw_get_device_info(int devid, int revision);
+const char *brw_get_device_name(int devid);
diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp
index b2701b89689..8a9bbdbae52 100644
--- a/src/mesa/drivers/dri/i965/brw_fs.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs.cpp
@@ -1910,6 +1910,10 @@ fs_visitor::assign_vs_urb_setup()
    unsigned vue_entries =
       MAX2(count, vs_prog_data->base.vue_map.num_slots);
 
+   /* URB entry size is counted in units of 64 bytes (for the 3DSTATE_URB_VS
+    * command).  Each attribute is 16 bytes (4 floats/dwords), so each unit
+    * fits four attributes.
+    */
    vs_prog_data->base.urb_entry_size = ALIGN(vue_entries, 4) / 4;
    vs_prog_data->base.urb_read_length = (count + 1) / 2;
 
@@ -3033,9 +3037,22 @@ fs_visitor::emit_repclear_shader()
    brw_wm_prog_key *key = (brw_wm_prog_key*) this->key;
    int base_mrf = 1;
    int color_mrf = base_mrf + 2;
+   fs_inst *mov;
+
+   if (uniforms == 1) {
+      mov = emit(MOV(vec4(brw_message_reg(color_mrf)),
+                     fs_reg(UNIFORM, 0, BRW_REGISTER_TYPE_F)));
+   } else {
+      struct brw_reg reg =
+         brw_reg(BRW_GENERAL_REGISTER_FILE,
+                 2, 3, 0, 0, BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_8,
+                 BRW_WIDTH_2,
+                 BRW_HORIZONTAL_STRIDE_4, BRW_SWIZZLE_XYZW, WRITEMASK_XYZW);
+
+      mov = emit(MOV(vec4(brw_message_reg(color_mrf)), fs_reg(reg)));
+   }
 
-   fs_inst *mov = emit(MOV(vec4(brw_message_reg(color_mrf)),
-                           fs_reg(UNIFORM, 0, BRW_REGISTER_TYPE_F)));
    mov->force_writemask_all = true;
 
    fs_inst *write;
@@ -3065,8 +3082,10 @@ fs_visitor::emit_repclear_shader()
    assign_curb_setup();
 
    /* Now that we have the uniform assigned, go ahead and force it to a vec4. */
-   assert(mov->src[0].file == HW_REG);
-   mov->src[0] = brw_vec4_grf(mov->src[0].fixed_hw_reg.nr, 0);
+   if (uniforms == 1) {
+      assert(mov->src[0].file == HW_REG);
+      mov->src[0] = brw_vec4_grf(mov->src[0].fixed_hw_reg.nr, 0);
+   }
 }
 
 /**
@@ -4081,7 +4100,8 @@ fs_visitor::run_vs()
 {
    assert(stage == MESA_SHADER_VERTEX);
 
-   assign_common_binding_table_offsets(0);
+   if (prog_data->map_entries == NULL)
+      assign_common_binding_table_offsets(0);
    setup_vs_payload();
 
    if (INTEL_DEBUG & DEBUG_SHADER_TIME)
@@ -4129,7 +4149,8 @@ fs_visitor::run_fs()
 
    sanity_param_count = prog->Parameters->NumParameters;
 
-   assign_binding_table_offsets();
+   if (prog_data->map_entries == NULL)
+      assign_binding_table_offsets();
 
    if (devinfo->gen >= 6)
       setup_payload_gen6();
diff --git a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
index 80ca1b750f8..975f5f6b2c9 100644
--- a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
+++ b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
@@ -1198,14 +1198,20 @@ fs_visitor::visit(ir_expression *ir)
       ir_constant *const_uniform_block = ir->operands[0]->as_constant();
       ir_constant *const_offset = ir->operands[1]->as_constant();
       fs_reg surf_index;
+      uint32_t binding, set, index, set_index;
 
       if (const_uniform_block) {
          /* The block index is a constant, so just emit the binding table entry
           * as an immediate.
           */
-         surf_index = fs_reg(stage_prog_data->binding_table.ubo_start +
-                                 const_uniform_block->value.u[0]);
+         index = const_uniform_block->value.u[0];
+         set = shader->base.UniformBlocks[index].Set;
+         set_index = shader->base.UniformBlocks[index].Index;
+         binding = stage_prog_data->bind_map[set][set_index];
+         surf_index = fs_reg(binding);
       } else {
+         assert(0 && "need more info from the ir for this.");
+
          /* The block index is not a constant. Evaluate the index expression
           * per-channel and add the base UBO index; we have to select a value
           * from any live channel.
@@ -2289,8 +2295,13 @@ fs_visitor::emit_texture(ir_texture_opcode op,
 void
 fs_visitor::visit(ir_texture *ir)
 {
-   uint32_t sampler =
-      _mesa_get_sampler_uniform_value(ir->sampler, shader_prog, prog);
+   uint32_t sampler;
+
+   ir_dereference_variable *deref_var = ir->sampler->as_dereference_variable();
+   assert(deref_var);
+   ir_variable *var = deref_var->var;
+
+   sampler = stage_prog_data->bind_map[var->data.set][var->data.index];
 
    ir_rvalue *nonconst_sampler_index =
       _mesa_get_sampler_array_nonconst_index(ir->sampler);
diff --git a/src/mesa/drivers/dri/i965/brw_gs.c b/src/mesa/drivers/dri/i965/brw_gs.c
index 52c73031a3c..45c132b4a9e 100644
--- a/src/mesa/drivers/dri/i965/brw_gs.c
+++ b/src/mesa/drivers/dri/i965/brw_gs.c
@@ -33,19 +33,23 @@
 #include "brw_state.h"
 #include "brw_ff_gs.h"
 
-
 bool
-brw_codegen_gs_prog(struct brw_context *brw,
+brw_compile_gs_prog(struct brw_context *brw,
                     struct gl_shader_program *prog,
                     struct brw_geometry_program *gp,
-                    struct brw_gs_prog_key *key)
+                    struct brw_gs_prog_key *key,
+                    struct brw_gs_compile_output *output)
 {
-   struct brw_stage_state *stage_state = &brw->gs.base;
    struct brw_gs_compile c;
    memset(&c, 0, sizeof(c));
    c.key = *key;
    c.gp = gp;
 
+   /* We get the bind map as input in the output struct...*/
+   c.prog_data.base.base.map_entries = output->prog_data.base.base.map_entries;
+   memcpy(c.prog_data.base.base.bind_map, output->prog_data.base.base.bind_map,
+          sizeof(c.prog_data.base.base.bind_map));
+
    c.prog_data.include_primitive_id =
       (gp->program.Base.InputsRead & VARYING_BIT_PRIMITIVE_ID) != 0;
 
@@ -274,18 +278,41 @@ brw_codegen_gs_prog(struct brw_context *brw,
 
       c.prog_data.base.base.total_scratch
          = brw_get_scratch_size(c.base.last_scratch*REG_SIZE);
+   }
+
+   output->mem_ctx = mem_ctx;
+   output->program = program;
+   output->program_size = program_size;
+   memcpy(&output->prog_data, &c.prog_data,
+          sizeof(output->prog_data));
+
+   return true;
+}
+
+bool
+brw_codegen_gs_prog(struct brw_context *brw,
+                    struct gl_shader_program *prog,
+                    struct brw_geometry_program *gp,
+                    struct brw_gs_prog_key *key)
+{
+   struct brw_gs_compile_output output;
+   struct brw_stage_state *stage_state = &brw->gs.base;
+
+   if (brw_compile_gs_prog(brw, prog, gp, key, &output))
+      return false;
 
+   if (output.prog_data.base.base.total_scratch) {
       brw_get_scratch_bo(brw, &stage_state->scratch_bo,
-			 c.prog_data.base.base.total_scratch *
+			 output.prog_data.base.base.total_scratch *
                          brw->max_gs_threads);
    }
 
    brw_upload_cache(&brw->cache, BRW_CACHE_GS_PROG,
-                    &c.key, sizeof(c.key),
-                    program, program_size,
-                    &c.prog_data, sizeof(c.prog_data),
+                    key, sizeof(*key),
+                    output.program, output.program_size,
+                    &output.prog_data, sizeof(output.prog_data),
                     &stage_state->prog_offset, &brw->gs.prog_data);
-   ralloc_free(mem_ctx);
+   ralloc_free(output.mem_ctx);
 
    return true;
 }
diff --git a/src/mesa/drivers/dri/i965/brw_gs.h b/src/mesa/drivers/dri/i965/brw_gs.h
index b327c40f140..573bbdb16f8 100644
--- a/src/mesa/drivers/dri/i965/brw_gs.h
+++ b/src/mesa/drivers/dri/i965/brw_gs.h
@@ -37,6 +37,22 @@ struct gl_context;
 struct gl_shader_program;
 struct gl_program;
 
+struct brw_gs_compile_output {
+   void *mem_ctx;
+   const void *program;
+   uint32_t program_size;
+   struct brw_gs_prog_data prog_data;
+};
+
+struct brw_gs_prog_key;
+
+bool
+brw_compile_gs_prog(struct brw_context *brw,
+                    struct gl_shader_program *prog,
+                    struct brw_geometry_program *gp,
+                    struct brw_gs_prog_key *key,
+                    struct brw_gs_compile_output *output);
+
 bool brw_gs_prog_data_compare(const void *a, const void *b);
 
 void
diff --git a/src/mesa/drivers/dri/i965/brw_state_cache.c b/src/mesa/drivers/dri/i965/brw_state_cache.c
index 157b33d4f4c..24778d25379 100644
--- a/src/mesa/drivers/dri/i965/brw_state_cache.c
+++ b/src/mesa/drivers/dri/i965/brw_state_cache.c
@@ -427,6 +427,9 @@ brw_destroy_cache(struct brw_context *brw, struct brw_cache *cache)
 
    DBG("%s\n", __func__);
 
+   if (cache->bo == NULL)
+      return;
+
    if (brw->has_llc)
       drm_intel_bo_unmap(cache->bo);
    drm_intel_bo_unreference(cache->bo);
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 45a03bba857..5496225a6c7 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -46,7 +46,7 @@
  * Return a bitfield where bit n is set if barycentric interpolation mode n
  * (see enum brw_wm_barycentric_interp_mode) is needed by the fragment shader.
  */
-static unsigned
+unsigned
 brw_compute_barycentric_interp_modes(struct brw_context *brw,
                                      bool shade_model_flat,
                                      bool persample_shading,
diff --git a/src/mesa/drivers/dri/i965/brw_wm.h b/src/mesa/drivers/dri/i965/brw_wm.h
index 0a8a97b2f5e..73a741f89e4 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.h
+++ b/src/mesa/drivers/dri/i965/brw_wm.h
@@ -90,6 +90,12 @@ bool brw_wm_prog_data_compare(const void *a, const void *b);
 void
 brw_upload_wm_prog(struct brw_context *brw);
 
+unsigned
+brw_compute_barycentric_interp_modes(struct brw_context *brw,
+                                     bool shade_model_flat,
+                                     bool persample_shading,
+                                     const struct gl_fragment_program *fprog);
+
 #ifdef __cplusplus
 } // extern "C"
 #endif
diff --git a/src/mesa/drivers/dri/i965/intel_debug.c b/src/mesa/drivers/dri/i965/intel_debug.c
index 53f575ab78f..33a0348486d 100644
--- a/src/mesa/drivers/dri/i965/intel_debug.c
+++ b/src/mesa/drivers/dri/i965/intel_debug.c
@@ -60,7 +60,7 @@ static const struct dri_debug_control debug_control[] = {
    { "urb",         DEBUG_URB },
    { "vs",          DEBUG_VS },
    { "clip",        DEBUG_CLIP },
-   { "aub",         DEBUG_AUB },
+   { "foob",        DEBUG_AUB }, /* disable aub dumbing in the dri driver */
    { "shader_time", DEBUG_SHADER_TIME },
    { "no16",        DEBUG_NO16 },
    { "blorp",       DEBUG_BLORP },
diff --git a/src/mesa/drivers/dri/i965/intel_screen.c b/src/mesa/drivers/dri/i965/intel_screen.c
index dda16389f8a..4860a160ee9 100644
--- a/src/mesa/drivers/dri/i965/intel_screen.c
+++ b/src/mesa/drivers/dri/i965/intel_screen.c
@@ -1416,6 +1416,78 @@ __DRIconfig **intelInitScreen2(__DRIscreen *psp)
    return (const __DRIconfig**) intel_screen_make_configs(psp);
 }
 
+struct intel_screen *
+intel_screen_create(int fd)
+{
+   __DRIscreen *psp;
+   __DRIconfig **configs;
+   int i;
+
+   psp = malloc(sizeof(*psp));
+   if (psp == NULL)
+      return NULL;
+
+   psp->image.loader = (void *) 1; /* Don't complain about this being NULL */
+   psp->fd = fd;
+   psp->dri2.useInvalidate = (void *) 1;
+
+   configs = (__DRIconfig **) intelInitScreen2(psp);
+   for (i = 0; configs[i]; i++)
+      free(configs[i]);
+   free(configs);
+
+   return psp->driverPrivate;
+}
+
+void
+intel_screen_destroy(struct intel_screen *screen)
+{
+   __DRIscreen *psp;
+
+   psp = screen->driScrnPriv;
+   intelDestroyScreen(screen->driScrnPriv);
+   free(psp);
+}
+
+
+struct brw_context *
+intel_context_create(struct intel_screen *screen)
+{
+   __DRIcontext *driContextPriv;
+   struct brw_context *brw;
+   unsigned error;
+
+   driContextPriv = malloc(sizeof(*driContextPriv));
+   if (driContextPriv == NULL)
+      return NULL;
+
+   driContextPriv->driScreenPriv = screen->driScrnPriv;
+
+   brwCreateContext(API_OPENGL_CORE,
+                    NULL, /* visual */
+                    driContextPriv,
+                    3, 0,
+                    0, /* flags */
+                    false, /* notify_reset */
+                    &error,
+                    NULL);
+
+   brw = driContextPriv->driverPrivate;
+   brw->ctx.FirstTimeCurrent = false;
+
+   return driContextPriv->driverPrivate;
+}
+
+void
+intel_context_destroy(struct brw_context *brw)
+{
+   __DRIcontext *driContextPriv;
+
+   driContextPriv = brw->driContext;
+   intelDestroyContext(driContextPriv);
+   free(driContextPriv);
+}
+
 struct intel_buffer {
    __DRIbuffer base;
    drm_intel_bo *bo;