1 /**************************************************************************
3 * Copyright 2007-2009 VMware, Inc.
6 * Permission is hereby granted, free of charge, to any person obtaining a
7 * copy of this software and associated documentation files (the
8 * "Software"), to deal in the Software without restriction, including
9 * without limitation the rights to use, copy, modify, merge, publish,
10 * distribute, sub license, and/or sell copies of the Software, and to
11 * permit persons to whom the Software is furnished to do so, subject to
12 * the following conditions:
14 * The above copyright notice and this permission notice (including the
15 * next paragraph) shall be included in all copies or substantial portions
18 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
19 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
20 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
21 * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
22 * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
23 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
24 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
26 **************************************************************************/
29 * Rasterization for binned triangles within a tile
33 #include "util/u_math.h"
36 #include "lp_rast_priv.h"
37 #include "lp_tile_soa.h"
43 * Shade all pixels in a 4x4 block.
46 block_full_4(struct lp_rasterizer_task
*task
,
47 const struct lp_rast_triangle
*tri
,
50 lp_rast_shade_quads_all(task
, &tri
->inputs
, x
, y
);
55 * Shade all pixels in a 16x16 block.
58 block_full_16(struct lp_rasterizer_task
*task
,
59 const struct lp_rast_triangle
*tri
,
65 for (iy
= 0; iy
< 16; iy
+= 4)
66 for (ix
= 0; ix
< 16; ix
+= 4)
67 block_full_4(task
, tri
, x
+ ix
, y
+ iy
);
71 static INLINE
unsigned
72 build_mask(int c
, int dcdx
, int dcdy
)
81 mask
|= ((c0
+ 0 * dcdy
) >> 31) & (1 << 0);
82 mask
|= ((c0
+ 1 * dcdy
) >> 31) & (1 << 2);
83 mask
|= ((c0
+ 2 * dcdy
) >> 31) & (1 << 8);
84 mask
|= ((c0
+ 3 * dcdy
) >> 31) & (1 << 10);
85 mask
|= ((c1
+ 0 * dcdy
) >> 31) & (1 << 1);
86 mask
|= ((c1
+ 1 * dcdy
) >> 31) & (1 << 3);
87 mask
|= ((c1
+ 2 * dcdy
) >> 31) & (1 << 9);
88 mask
|= ((c1
+ 3 * dcdy
) >> 31) & (1 << 11);
89 mask
|= ((c2
+ 0 * dcdy
) >> 31) & (1 << 4);
90 mask
|= ((c2
+ 1 * dcdy
) >> 31) & (1 << 6);
91 mask
|= ((c2
+ 2 * dcdy
) >> 31) & (1 << 12);
92 mask
|= ((c2
+ 3 * dcdy
) >> 31) & (1 << 14);
93 mask
|= ((c3
+ 0 * dcdy
) >> 31) & (1 << 5);
94 mask
|= ((c3
+ 1 * dcdy
) >> 31) & (1 << 7);
95 mask
|= ((c3
+ 2 * dcdy
) >> 31) & (1 << 13);
96 mask
|= ((c3
+ 3 * dcdy
) >> 31) & (1 << 15);
101 static INLINE
unsigned
102 build_mask_linear(int c
, int dcdx
, int dcdy
)
111 mask
|= ((c0
+ 0 * dcdx
) >> 31) & (1 << 0);
112 mask
|= ((c0
+ 1 * dcdx
) >> 31) & (1 << 1);
113 mask
|= ((c0
+ 2 * dcdx
) >> 31) & (1 << 2);
114 mask
|= ((c0
+ 3 * dcdx
) >> 31) & (1 << 3);
115 mask
|= ((c1
+ 0 * dcdx
) >> 31) & (1 << 4);
116 mask
|= ((c1
+ 1 * dcdx
) >> 31) & (1 << 5);
117 mask
|= ((c1
+ 2 * dcdx
) >> 31) & (1 << 6);
118 mask
|= ((c1
+ 3 * dcdx
) >> 31) & (1 << 7);
119 mask
|= ((c2
+ 0 * dcdx
) >> 31) & (1 << 8);
120 mask
|= ((c2
+ 1 * dcdx
) >> 31) & (1 << 9);
121 mask
|= ((c2
+ 2 * dcdx
) >> 31) & (1 << 10);
122 mask
|= ((c2
+ 3 * dcdx
) >> 31) & (1 << 11);
123 mask
|= ((c3
+ 0 * dcdx
) >> 31) & (1 << 12);
124 mask
|= ((c3
+ 1 * dcdx
) >> 31) & (1 << 13);
125 mask
|= ((c3
+ 2 * dcdx
) >> 31) & (1 << 14);
126 mask
|= ((c3
+ 3 * dcdx
) >> 31) & (1 << 15);
134 #include "lp_rast_tri_tmp.h"
138 #include "lp_rast_tri_tmp.h"
142 #include "lp_rast_tri_tmp.h"
146 #include "lp_rast_tri_tmp.h"
150 #include "lp_rast_tri_tmp.h"
154 #include "lp_rast_tri_tmp.h"
158 #include "lp_rast_tri_tmp.h"
161 /* Special case for 3 plane triangle which is contained entirely
162 * within a 16x16 block.
165 lp_rast_triangle_3_16(struct lp_rasterizer_task
*task
,
166 const union lp_rast_cmd_arg arg
)
168 const struct lp_rast_triangle
*tri
= arg
.triangle
.tri
;
169 const struct lp_rast_plane
*plane
= tri
->plane
;
170 unsigned mask
= arg
.triangle
.plane_mask
;
171 const int x
= task
->x
+ (mask
& 0xf) * 16;
172 const int y
= task
->y
+ (mask
>> 4) * 16;
173 unsigned outmask
, inmask
, partmask
, partial_mask
;
177 outmask
= 0; /* outside one or more trivial reject planes */
178 partmask
= 0; /* outside one or more trivial accept planes */
180 for (j
= 0; j
< 3; j
++) {
181 c
[j
] = plane
[j
].c
+ plane
[j
].dcdy
* y
- plane
[j
].dcdx
* x
;
184 const int dcdx
= -plane
[j
].dcdx
* 4;
185 const int dcdy
= plane
[j
].dcdy
* 4;
186 const int cox
= c
[j
] + plane
[j
].eo
* 4;
187 const int cio
= c
[j
] + plane
[j
].ei
* 4 - 1;
189 outmask
|= build_mask_linear(cox
, dcdx
, dcdy
);
190 partmask
|= build_mask_linear(cio
, dcdx
, dcdy
);
194 if (outmask
== 0xffff)
197 /* Mask of sub-blocks which are inside all trivial accept planes:
199 inmask
= ~partmask
& 0xffff;
201 /* Mask of sub-blocks which are inside all trivial reject planes,
202 * but outside at least one trivial accept plane:
204 partial_mask
= partmask
& ~outmask
;
206 assert((partial_mask
& inmask
) == 0);
208 /* Iterate over partials:
210 while (partial_mask
) {
211 int i
= ffs(partial_mask
) - 1;
212 int ix
= (i
& 3) * 4;
213 int iy
= (i
>> 2) * 4;
218 partial_mask
&= ~(1 << i
);
220 for (j
= 0; j
< 3; j
++)
223 + plane
[j
].dcdy
* iy
);
225 do_block_4_3(task
, tri
, plane
, px
, py
, cx
);
228 /* Iterate over fulls:
231 int i
= ffs(inmask
) - 1;
232 int ix
= (i
& 3) * 4;
233 int iy
= (i
>> 2) * 4;
239 block_full_4(task
, tri
, px
, py
);