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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
|
/*
* Copyright © 2011 Intel Corporation
*
* 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 AUTHORS OR COPYRIGHT HOLDERS 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.
*/
#include <stdbool.h>
#include "brw_context.h"
#include "brw_state.h"
#include "brw_defines.h"
#include "brw_util.h"
#include "brw_wm.h"
#include "program/program.h"
#include "program/prog_parameter.h"
#include "program/prog_statevars.h"
#include "intel_batchbuffer.h"
static void
upload_wm_state(struct brw_context *brw)
{
struct gl_context *ctx = &brw->ctx;
/* BRW_NEW_FRAGMENT_PROGRAM */
const struct brw_fragment_program *fp =
brw_fragment_program_const(brw->fragment_program);
/* BRW_NEW_FS_PROG_DATA */
const struct brw_wm_prog_data *prog_data = brw->wm.prog_data;
bool writes_depth = prog_data->computed_depth_mode != BRW_PSCDEPTH_OFF;
uint32_t dw1, dw2;
/* _NEW_BUFFERS */
bool multisampled_fbo = ctx->DrawBuffer->Visual.samples > 1;
dw1 = dw2 = 0;
dw1 |= GEN7_WM_STATISTICS_ENABLE;
dw1 |= GEN7_WM_LINE_AA_WIDTH_1_0;
dw1 |= GEN7_WM_LINE_END_CAP_AA_WIDTH_0_5;
/* _NEW_LINE */
if (ctx->Line.StippleFlag)
dw1 |= GEN7_WM_LINE_STIPPLE_ENABLE;
/* _NEW_POLYGON */
if (ctx->Polygon.StippleFlag)
dw1 |= GEN7_WM_POLYGON_STIPPLE_ENABLE;
if (fp->program.Base.InputsRead & VARYING_BIT_POS)
dw1 |= GEN7_WM_USES_SOURCE_DEPTH | GEN7_WM_USES_SOURCE_W;
dw1 |= prog_data->computed_depth_mode << GEN7_WM_COMPUTED_DEPTH_MODE_SHIFT;
dw1 |= prog_data->barycentric_interp_modes <<
GEN7_WM_BARYCENTRIC_INTERPOLATION_MODE_SHIFT;
/* _NEW_COLOR, _NEW_MULTISAMPLE */
/* Enable if the pixel shader kernel generates and outputs oMask.
*/
if (prog_data->uses_kill || ctx->Color.AlphaEnabled ||
ctx->Multisample.SampleAlphaToCoverage ||
prog_data->uses_omask) {
dw1 |= GEN7_WM_KILL_ENABLE;
}
/* _NEW_BUFFERS | _NEW_COLOR */
if (brw_color_buffer_write_enabled(brw) || writes_depth ||
dw1 & GEN7_WM_KILL_ENABLE) {
dw1 |= GEN7_WM_DISPATCH_ENABLE;
}
if (multisampled_fbo) {
/* _NEW_MULTISAMPLE */
if (ctx->Multisample.Enabled)
dw1 |= GEN7_WM_MSRAST_ON_PATTERN;
else
dw1 |= GEN7_WM_MSRAST_OFF_PIXEL;
if (_mesa_get_min_invocations_per_fragment(ctx, brw->fragment_program, false) > 1)
dw2 |= GEN7_WM_MSDISPMODE_PERSAMPLE;
else
dw2 |= GEN7_WM_MSDISPMODE_PERPIXEL;
} else {
dw1 |= GEN7_WM_MSRAST_OFF_PIXEL;
dw2 |= GEN7_WM_MSDISPMODE_PERSAMPLE;
}
if (fp->program.Base.SystemValuesRead & SYSTEM_BIT_SAMPLE_MASK_IN) {
dw1 |= GEN7_WM_USES_INPUT_COVERAGE_MASK;
}
BEGIN_BATCH(3);
OUT_BATCH(_3DSTATE_WM << 16 | (3 - 2));
OUT_BATCH(dw1);
OUT_BATCH(dw2);
ADVANCE_BATCH();
}
const struct brw_tracked_state gen7_wm_state = {
.dirty = {
.mesa = _NEW_BUFFERS |
_NEW_COLOR |
_NEW_LINE |
_NEW_MULTISAMPLE |
_NEW_POLYGON,
.brw = BRW_NEW_BATCH |
BRW_NEW_FRAGMENT_PROGRAM |
BRW_NEW_FS_PROG_DATA,
},
.emit = upload_wm_state,
};
static void
upload_ps_state(struct brw_context *brw)
{
struct gl_context *ctx = &brw->ctx;
uint32_t dw2, dw4, dw5, ksp0, ksp2;
const int max_threads_shift = brw->is_haswell ?
HSW_PS_MAX_THREADS_SHIFT : IVB_PS_MAX_THREADS_SHIFT;
/* BRW_NEW_FS_PROG_DATA */
const struct brw_wm_prog_data *prog_data = brw->wm.prog_data;
dw2 = dw4 = dw5 = ksp2 = 0;
dw2 |=
(ALIGN(brw->wm.base.sampler_count, 4) / 4) << GEN7_PS_SAMPLER_COUNT_SHIFT;
dw2 |= ((prog_data->base.binding_table.size_bytes / 4) <<
GEN7_PS_BINDING_TABLE_ENTRY_COUNT_SHIFT);
if (prog_data->base.use_alt_mode)
dw2 |= GEN7_PS_FLOATING_POINT_MODE_ALT;
/* Haswell requires the sample mask to be set in this packet as well as
* in 3DSTATE_SAMPLE_MASK; the values should match. */
/* _NEW_BUFFERS, _NEW_MULTISAMPLE */
if (brw->is_haswell)
dw4 |= SET_FIELD(gen6_determine_sample_mask(brw), HSW_PS_SAMPLE_MASK);
dw4 |= (brw->max_wm_threads - 1) << max_threads_shift;
if (prog_data->base.nr_params > 0)
dw4 |= GEN7_PS_PUSH_CONSTANT_ENABLE;
/* From the IVB PRM, volume 2 part 1, page 287:
* "This bit is inserted in the PS payload header and made available to
* the DataPort (either via the message header or via header bypass) to
* indicate that oMask data (one or two phases) is included in Render
* Target Write messages. If present, the oMask data is used to mask off
* samples."
*/
if (prog_data->uses_omask)
dw4 |= GEN7_PS_OMASK_TO_RENDER_TARGET;
/* From the IVB PRM, volume 2 part 1, page 287:
* "If the PS kernel does not need the Position XY Offsets to
* compute a Position Value, then this field should be programmed
* to POSOFFSET_NONE."
* "SW Recommendation: If the PS kernel needs the Position Offsets
* to compute a Position XY value, this field should match Position
* ZW Interpolation Mode to ensure a consistent position.xyzw
* computation."
* We only require XY sample offsets. So, this recommendation doesn't
* look useful at the moment. We might need this in future.
*/
if (prog_data->uses_pos_offset)
dw4 |= GEN7_PS_POSOFFSET_SAMPLE;
else
dw4 |= GEN7_PS_POSOFFSET_NONE;
/* BRW_NEW_FS_PROG_DATA | _NEW_COLOR
*
* The hardware wedges if you have this bit set but don't turn on any dual
* source blend factors.
*/
if (prog_data->dual_src_blend &&
(ctx->Color.BlendEnabled & 1) &&
ctx->Color.Blend[0]._UsesDualSrc) {
dw4 |= GEN7_PS_DUAL_SOURCE_BLEND_ENABLE;
}
/* BRW_NEW_FS_PROG_DATA */
if (prog_data->num_varying_inputs != 0)
dw4 |= GEN7_PS_ATTRIBUTE_ENABLE;
/* In case of non 1x per sample shading, only one of SIMD8 and SIMD16
* should be enabled. We do 'SIMD16 only' dispatch if a SIMD16 shader
* is successfully compiled. In majority of the cases that bring us
* better performance than 'SIMD8 only' dispatch.
*/
int min_inv_per_frag =
_mesa_get_min_invocations_per_fragment(ctx, brw->fragment_program, false);
assert(min_inv_per_frag >= 1);
if (prog_data->prog_offset_16 || prog_data->no_8) {
dw4 |= GEN7_PS_16_DISPATCH_ENABLE;
if (!prog_data->no_8 && min_inv_per_frag == 1) {
dw4 |= GEN7_PS_8_DISPATCH_ENABLE;
dw5 |= (prog_data->base.dispatch_grf_start_reg <<
GEN7_PS_DISPATCH_START_GRF_SHIFT_0);
dw5 |= (prog_data->dispatch_grf_start_reg_16 <<
GEN7_PS_DISPATCH_START_GRF_SHIFT_2);
ksp0 = brw->wm.base.prog_offset;
ksp2 = brw->wm.base.prog_offset + prog_data->prog_offset_16;
} else {
dw5 |= (prog_data->dispatch_grf_start_reg_16 <<
GEN7_PS_DISPATCH_START_GRF_SHIFT_0);
ksp0 = brw->wm.base.prog_offset + prog_data->prog_offset_16;
}
}
else {
dw4 |= GEN7_PS_8_DISPATCH_ENABLE;
dw5 |= (prog_data->base.dispatch_grf_start_reg <<
GEN7_PS_DISPATCH_START_GRF_SHIFT_0);
ksp0 = brw->wm.base.prog_offset;
}
dw4 |= brw->wm.fast_clear_op;
BEGIN_BATCH(8);
OUT_BATCH(_3DSTATE_PS << 16 | (8 - 2));
OUT_BATCH(ksp0);
OUT_BATCH(dw2);
if (prog_data->base.total_scratch) {
OUT_RELOC(brw->wm.base.scratch_bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
ffs(prog_data->base.total_scratch) - 11);
} else {
OUT_BATCH(0);
}
OUT_BATCH(dw4);
OUT_BATCH(dw5);
OUT_BATCH(0); /* kernel 1 pointer */
OUT_BATCH(ksp2);
ADVANCE_BATCH();
}
const struct brw_tracked_state gen7_ps_state = {
.dirty = {
.mesa = _NEW_BUFFERS |
_NEW_COLOR |
_NEW_MULTISAMPLE,
.brw = BRW_NEW_BATCH |
BRW_NEW_FRAGMENT_PROGRAM |
BRW_NEW_FS_PROG_DATA,
},
.emit = upload_ps_state,
};
|
