1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
|
/*
Copyright (C) Intel Corp. 2006. All Rights Reserved.
Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
develop this 3D driver.
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice (including the
next paragraph) shall be included in all copies or substantial
portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
**********************************************************************/
/*
* Authors:
* Keith Whitwell <keith@tungstengraphics.com>
*/
#include "main/glheader.h"
#include "main/macros.h"
#include "main/enums.h"
#include "program/program.h"
#include "intel_batchbuffer.h"
#include "brw_defines.h"
#include "brw_context.h"
#include "brw_eu.h"
#include "brw_gs.h"
static void brw_gs_alloc_regs( struct brw_gs_compile *c,
GLuint nr_verts )
{
GLuint i = 0,j;
/* Register usage is static, precompute here:
*/
c->reg.R0 = retype(brw_vec8_grf(i, 0), BRW_REGISTER_TYPE_UD); i++;
/* Payload vertices plus space for more generated vertices:
*/
for (j = 0; j < nr_verts; j++) {
c->reg.vertex[j] = brw_vec4_grf(i, 0);
i += c->nr_regs;
}
c->reg.temp = brw_vec8_grf(i, 0);
c->prog_data.urb_read_length = c->nr_regs;
c->prog_data.total_grf = i;
}
static void brw_gs_emit_vue(struct brw_gs_compile *c,
struct brw_reg vert,
bool last,
GLuint header)
{
struct brw_compile *p = &c->func;
struct intel_context *intel = &c->func.brw->intel;
bool allocate = !last;
struct brw_reg temp;
if (intel->gen < 6)
temp = c->reg.R0;
else {
temp = c->reg.temp;
brw_MOV(p, retype(temp, BRW_REGISTER_TYPE_UD),
retype(c->reg.R0, BRW_REGISTER_TYPE_UD));
}
/* Overwrite PrimType and PrimStart in the message header, for
* each vertex in turn:
*/
brw_MOV(p, get_element_ud(temp, 2), brw_imm_ud(header));
/* Copy the vertex from vertn into m1..mN+1:
*/
brw_copy8(p, brw_message_reg(1), vert, c->nr_regs);
/* Send each vertex as a seperate write to the urb. This is
* different to the concept in brw_sf_emit.c, where subsequent
* writes are used to build up a single urb entry. Each of these
* writes instantiates a seperate urb entry, and a new one must be
* allocated each time.
*/
brw_urb_WRITE(p,
allocate ? temp : retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
0,
temp,
allocate,
1, /* used */
c->nr_regs + 1, /* msg length */
allocate ? 1 : 0, /* response length */
allocate ? 0 : 1, /* eot */
1, /* writes_complete */
0, /* urb offset */
BRW_URB_SWIZZLE_NONE);
if (intel->gen >= 6 && allocate)
brw_MOV(p, get_element_ud(c->reg.R0, 0), get_element_ud(temp, 0));
}
static void brw_gs_ff_sync(struct brw_gs_compile *c, int num_prim)
{
struct brw_compile *p = &c->func;
struct intel_context *intel = &c->func.brw->intel;
if (intel->gen < 6) {
brw_MOV(p, get_element_ud(c->reg.R0, 1), brw_imm_ud(num_prim));
brw_ff_sync(p,
c->reg.R0,
0,
c->reg.R0,
1, /* allocate */
1, /* response length */
0 /* eot */);
} else {
brw_MOV(p, retype(c->reg.temp, BRW_REGISTER_TYPE_UD),
retype(c->reg.R0, BRW_REGISTER_TYPE_UD));
brw_MOV(p, get_element_ud(c->reg.temp, 1), brw_imm_ud(num_prim));
brw_ff_sync(p,
c->reg.temp,
0,
c->reg.temp,
1, /* allocate */
1, /* response length */
0 /* eot */);
brw_MOV(p, get_element_ud(c->reg.R0, 0),
get_element_ud(c->reg.temp, 0));
}
}
void brw_gs_quads( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
{
struct intel_context *intel = &c->func.brw->intel;
brw_gs_alloc_regs(c, 4);
/* Use polygons for correct edgeflag behaviour. Note that vertex 3
* is the PV for quads, but vertex 0 for polygons:
*/
if (intel->needs_ff_sync)
brw_gs_ff_sync(c, 1);
if (key->pv_first) {
brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[2], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[3], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
}
else {
brw_gs_emit_vue(c, c->reg.vertex[3], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
brw_gs_emit_vue(c, c->reg.vertex[0], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[2], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
}
}
void brw_gs_quad_strip( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
{
struct intel_context *intel = &c->func.brw->intel;
brw_gs_alloc_regs(c, 4);
if (intel->needs_ff_sync)
brw_gs_ff_sync(c, 1);
if (key->pv_first) {
brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[2], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[3], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
}
else {
brw_gs_emit_vue(c, c->reg.vertex[2], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
brw_gs_emit_vue(c, c->reg.vertex[3], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[0], 0, (_3DPRIM_POLYGON << 2));
brw_gs_emit_vue(c, c->reg.vertex[1], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
}
}
void brw_gs_lines( struct brw_gs_compile *c )
{
struct intel_context *intel = &c->func.brw->intel;
brw_gs_alloc_regs(c, 2);
if (intel->needs_ff_sync)
brw_gs_ff_sync(c, 1);
brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_START));
brw_gs_emit_vue(c, c->reg.vertex[1], 1, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_END));
}
|
