X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;f=src%2Fgallium%2Fdrivers%2Fllvmpipe%2Flp_rast.h;h=c209f47f0f5e2e82841c42fa10aac8ccc136e3d7;hb=00faf9f00059370bc20ebeaf00884c2d8ef15a74;hp=a64c152cf8317164292e7eee36475a8f444bcb6e;hpb=0072acd447dc6be652e63752e50215c3105322c8;p=mesa.git

diff --git a/src/gallium/drivers/llvmpipe/lp_rast.h b/src/gallium/drivers/llvmpipe/lp_rast.h
index a64c152cf83..c209f47f0f5 100644
--- a/src/gallium/drivers/llvmpipe/lp_rast.h
+++ b/src/gallium/drivers/llvmpipe/lp_rast.h
@@ -38,6 +38,7 @@
 #define LP_RAST_H
 
 #include "pipe/p_compiler.h"
+#include "util/u_pack_color.h"
 #include "lp_jit.h"
 
 
@@ -46,10 +47,28 @@ struct lp_scene;
 struct lp_fence;
 struct cmd_bin;
 
+#define FIXED_TYPE_WIDTH 64
 /** For sub-pixel positioning */
-#define FIXED_ORDER 4
+#define FIXED_ORDER 8
 #define FIXED_ONE (1<<FIXED_ORDER)
+#define FIXED_SHIFT (FIXED_TYPE_WIDTH - 1)
+/** Maximum length of an edge in a primitive in pixels.
+ *  If the framebuffer is large we have to think about fixed-point
+ *  integer overflow. Coordinates need ((FIXED_TYPE_WIDTH/2) - 1) bits
+ *  to be able to fit product of two such coordinates inside
+ *  FIXED_TYPE_WIDTH, any larger and we could overflow a
+ *  FIXED_TYPE_WIDTH_-bit int.
+ */
+#define MAX_FIXED_LENGTH (1 << (((FIXED_TYPE_WIDTH/2) - 1) - FIXED_ORDER))
+
+#define MAX_FIXED_LENGTH32 (1 << (((32/2) - 1) - FIXED_ORDER))
+
+/* Rasterizer output size going to jit fs, width/height */
+#define LP_RASTER_BLOCK_SIZE 4
+
+#define LP_MAX_ACTIVE_BINNED_QUERIES 64
 
+#define IMUL64(a, b) (((int64_t)(a)) * ((int64_t)(b)))
 
 struct lp_rasterizer_task;
 
@@ -83,23 +102,20 @@ struct lp_rast_shader_inputs {
    unsigned opaque:1;           /** Is opaque */
    unsigned pad0:29;            /* wasted space */
    unsigned stride;             /* how much to advance data between a0, dadx, dady */
-   unsigned pad2;               /* wasted space */
-   unsigned pad3;               /* wasted space */
+   unsigned layer;              /* the layer to render to (from gs, already clamped) */
+   unsigned viewport_index;     /* the active viewport index (from gs, already clamped) */
    /* followed by a0, dadx, dady and planes[] */
 };
 
-/* Note: the order of these values is important as they are loaded by
- * sse code in rasterization:
- */
 struct lp_rast_plane {
    /* edge function values at minx,miny ?? */
-   int c;
+   int64_t c;
 
-   int dcdx;
-   int dcdy;
+   int32_t dcdx;
+   int32_t dcdy;
 
    /* one-pixel sized trivial reject offsets for each plane */
-   int eo;
+   int64_t eo;
 };
 
 /**
@@ -121,6 +137,12 @@ struct lp_rast_triangle {
 };
 
 
+struct lp_rast_clear_rb {
+   union util_color color_val;
+   unsigned cbuf;
+};
+
+
 #define GET_A0(inputs) ((float (*)[4])((inputs)+1))
 #define GET_DADX(inputs) ((float (*)[4])((char *)((inputs) + 1) + (inputs)->stride))
 #define GET_DADY(inputs) ((float (*)[4])((char *)((inputs) + 1) + 2 * (inputs)->stride))
@@ -134,9 +156,6 @@ lp_rast_create( unsigned num_threads );
 void
 lp_rast_destroy( struct lp_rasterizer * );
 
-unsigned
-lp_rast_get_num_threads( struct lp_rasterizer * );
-
 void 
 lp_rast_queue_scene( struct lp_rasterizer *rast,
                      struct lp_scene *scene );
@@ -152,10 +171,10 @@ union lp_rast_cmd_arg {
       unsigned plane_mask;
    } triangle;
    const struct lp_rast_state *set_state;
-   uint8_t clear_color[4];
+   const struct lp_rast_clear_rb *clear_rb;
    struct {
-      uint32_t value;
-      uint32_t mask;
+      uint64_t value;
+      uint64_t mask;
    } clear_zstencil;
    const struct lp_rast_state *state;
    struct lp_fence *fence;
@@ -183,6 +202,22 @@ lp_rast_arg_triangle( const struct lp_rast_triangle *triangle,
    return arg;
 }
 
+/**
+ * Build argument for a contained triangle.
+ *
+ * All planes are enabled, so instead of the plane mask we pass the upper
+ * left coordinates of the a block that fully encloses the triangle.
+ */
+static INLINE union lp_rast_cmd_arg
+lp_rast_arg_triangle_contained( const struct lp_rast_triangle *triangle,
+                                unsigned x, unsigned y)
+{
+   union lp_rast_cmd_arg arg;
+   arg.triangle.tri = triangle;
+   arg.triangle.plane_mask = x | (y << 8);
+   return arg;
+}
+
 static INLINE union lp_rast_cmd_arg
 lp_rast_arg_state( const struct lp_rast_state *state )
 {
@@ -201,7 +236,7 @@ lp_rast_arg_fence( struct lp_fence *fence )
 
 
 static INLINE union lp_rast_cmd_arg
-lp_rast_arg_clearzs( unsigned value, unsigned mask )
+lp_rast_arg_clearzs( uint64_t value, uint64_t mask )
 {
    union lp_rast_cmd_arg arg;
    arg.clear_zstencil.value = value;
@@ -250,8 +285,19 @@ lp_rast_arg_null( void )
 #define LP_RAST_OP_BEGIN_QUERY       0xf
 #define LP_RAST_OP_END_QUERY         0x10
 #define LP_RAST_OP_SET_STATE         0x11
-
-#define LP_RAST_OP_MAX               0x12
+#define LP_RAST_OP_TRIANGLE_32_1     0x12
+#define LP_RAST_OP_TRIANGLE_32_2     0x13
+#define LP_RAST_OP_TRIANGLE_32_3     0x14
+#define LP_RAST_OP_TRIANGLE_32_4     0x15
+#define LP_RAST_OP_TRIANGLE_32_5     0x16
+#define LP_RAST_OP_TRIANGLE_32_6     0x17
+#define LP_RAST_OP_TRIANGLE_32_7     0x18
+#define LP_RAST_OP_TRIANGLE_32_8     0x19
+#define LP_RAST_OP_TRIANGLE_32_3_4   0x1a
+#define LP_RAST_OP_TRIANGLE_32_3_16  0x1b
+#define LP_RAST_OP_TRIANGLE_32_4_16  0x1c
+
+#define LP_RAST_OP_MAX               0x1d
 #define LP_RAST_OP_MASK              0xff
 
 void
@@ -262,4 +308,17 @@ void
 lp_debug_draw_bins_by_coverage( struct lp_scene *scene );
 
 
+#ifdef PIPE_ARCH_SSE
+#include <emmintrin.h>
+#include "util/u_sse.h"
+
+static INLINE __m128i
+lp_plane_to_m128i(const struct lp_rast_plane *plane)
+{
+   return _mm_setr_epi32((int32_t)plane->c, (int32_t)plane->dcdx,
+                         (int32_t)plane->dcdy, (int32_t)plane->eo);
+}
+
+#endif
+
 #endif