i965/blorp: Add an is_render_target parameter to surface_info::set.
[mesa.git] / src / mesa / drivers / dri / i965 / gen7_wm_state.c
1 /*
2 * Copyright © 2011 Intel Corporation
3 *
4 * Permission is hereby granted, free of charge, to any person obtaining a
5 * copy of this software and associated documentation files (the "Software"),
6 * to deal in the Software without restriction, including without limitation
7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
8 * and/or sell copies of the Software, and to permit persons to whom the
9 * Software is furnished to do so, subject to the following conditions:
10 *
11 * The above copyright notice and this permission notice (including the next
12 * paragraph) shall be included in all copies or substantial portions of the
13 * Software.
14 *
15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
18 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
20 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
21 * IN THE SOFTWARE.
22 */
23
24 #include <stdbool.h>
25 #include "brw_context.h"
26 #include "brw_state.h"
27 #include "brw_defines.h"
28 #include "brw_util.h"
29 #include "brw_wm.h"
30 #include "program/prog_parameter.h"
31 #include "program/prog_statevars.h"
32 #include "intel_batchbuffer.h"
33
34 static void
35 upload_wm_state(struct brw_context *brw)
36 {
37 struct gl_context *ctx = &brw->ctx;
38 const struct brw_fragment_program *fp =
39 brw_fragment_program_const(brw->fragment_program);
40 bool writes_depth = false;
41 uint32_t dw1, dw2;
42
43 /* _NEW_BUFFERS */
44 bool multisampled_fbo = ctx->DrawBuffer->Visual.samples > 1;
45
46 dw1 = dw2 = 0;
47 dw1 |= GEN7_WM_STATISTICS_ENABLE;
48 dw1 |= GEN7_WM_LINE_AA_WIDTH_1_0;
49 dw1 |= GEN7_WM_LINE_END_CAP_AA_WIDTH_0_5;
50
51 /* _NEW_LINE */
52 if (ctx->Line.StippleFlag)
53 dw1 |= GEN7_WM_LINE_STIPPLE_ENABLE;
54
55 /* _NEW_POLYGON */
56 if (ctx->Polygon.StippleFlag)
57 dw1 |= GEN7_WM_POLYGON_STIPPLE_ENABLE;
58
59 /* BRW_NEW_FRAGMENT_PROGRAM */
60 if (fp->program.Base.InputsRead & VARYING_BIT_POS)
61 dw1 |= GEN7_WM_USES_SOURCE_DEPTH | GEN7_WM_USES_SOURCE_W;
62 if (fp->program.Base.OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
63 writes_depth = fp->program.FragDepthLayout != FRAG_DEPTH_LAYOUT_UNCHANGED;
64
65 switch (fp->program.FragDepthLayout) {
66 case FRAG_DEPTH_LAYOUT_NONE:
67 case FRAG_DEPTH_LAYOUT_ANY:
68 dw1 |= GEN7_WM_PSCDEPTH_ON;
69 break;
70 case FRAG_DEPTH_LAYOUT_GREATER:
71 dw1 |= GEN7_WM_PSCDEPTH_ON_GE;
72 break;
73 case FRAG_DEPTH_LAYOUT_LESS:
74 dw1 |= GEN7_WM_PSCDEPTH_ON_LE;
75 break;
76 case FRAG_DEPTH_LAYOUT_UNCHANGED:
77 break;
78 }
79 }
80 /* CACHE_NEW_WM_PROG */
81 dw1 |= brw->wm.prog_data->barycentric_interp_modes <<
82 GEN7_WM_BARYCENTRIC_INTERPOLATION_MODE_SHIFT;
83
84 /* _NEW_COLOR, _NEW_MULTISAMPLE */
85 if (fp->program.UsesKill || ctx->Color.AlphaEnabled ||
86 ctx->Multisample.SampleAlphaToCoverage)
87 dw1 |= GEN7_WM_KILL_ENABLE;
88
89 /* _NEW_BUFFERS */
90 if (brw_color_buffer_write_enabled(brw) || writes_depth ||
91 dw1 & GEN7_WM_KILL_ENABLE) {
92 dw1 |= GEN7_WM_DISPATCH_ENABLE;
93 }
94 if (multisampled_fbo) {
95 /* _NEW_MULTISAMPLE */
96 if (ctx->Multisample.Enabled)
97 dw1 |= GEN7_WM_MSRAST_ON_PATTERN;
98 else
99 dw1 |= GEN7_WM_MSRAST_OFF_PIXEL;
100 dw2 |= GEN7_WM_MSDISPMODE_PERPIXEL;
101 } else {
102 dw1 |= GEN7_WM_MSRAST_OFF_PIXEL;
103 dw2 |= GEN7_WM_MSDISPMODE_PERSAMPLE;
104 }
105
106 BEGIN_BATCH(3);
107 OUT_BATCH(_3DSTATE_WM << 16 | (3 - 2));
108 OUT_BATCH(dw1);
109 OUT_BATCH(dw2);
110 ADVANCE_BATCH();
111 }
112
113 const struct brw_tracked_state gen7_wm_state = {
114 .dirty = {
115 .mesa = (_NEW_LINE | _NEW_POLYGON |
116 _NEW_COLOR | _NEW_BUFFERS |
117 _NEW_MULTISAMPLE),
118 .brw = (BRW_NEW_FRAGMENT_PROGRAM |
119 BRW_NEW_BATCH),
120 .cache = CACHE_NEW_WM_PROG,
121 },
122 .emit = upload_wm_state,
123 };
124
125 static void
126 upload_ps_state(struct brw_context *brw)
127 {
128 struct gl_context *ctx = &brw->ctx;
129 uint32_t dw2, dw4, dw5;
130 const int max_threads_shift = brw->is_haswell ?
131 HSW_PS_MAX_THREADS_SHIFT : IVB_PS_MAX_THREADS_SHIFT;
132
133 /* BRW_NEW_PS_BINDING_TABLE */
134 BEGIN_BATCH(2);
135 OUT_BATCH(_3DSTATE_BINDING_TABLE_POINTERS_PS << 16 | (2 - 2));
136 OUT_BATCH(brw->wm.base.bind_bo_offset);
137 ADVANCE_BATCH();
138
139 /* CACHE_NEW_SAMPLER */
140 BEGIN_BATCH(2);
141 OUT_BATCH(_3DSTATE_SAMPLER_STATE_POINTERS_PS << 16 | (2 - 2));
142 OUT_BATCH(brw->wm.base.sampler_offset);
143 ADVANCE_BATCH();
144
145 /* CACHE_NEW_WM_PROG */
146 gen7_upload_constant_state(brw, &brw->wm.base, true, _3DSTATE_CONSTANT_PS);
147
148 dw2 = dw4 = dw5 = 0;
149
150 /* CACHE_NEW_SAMPLER */
151 dw2 |=
152 (ALIGN(brw->wm.base.sampler_count, 4) / 4) << GEN7_PS_SAMPLER_COUNT_SHIFT;
153
154 /* Use ALT floating point mode for ARB fragment programs, because they
155 * require 0^0 == 1. Even though _CurrentFragmentProgram is used for
156 * rendering, CurrentFragmentProgram is used for this check to
157 * differentiate between the GLSL and non-GLSL cases.
158 */
159 /* BRW_NEW_FRAGMENT_PROGRAM */
160 if (ctx->Shader.CurrentFragmentProgram == NULL)
161 dw2 |= GEN7_PS_FLOATING_POINT_MODE_ALT;
162
163 if (brw->is_haswell)
164 dw4 |= SET_FIELD(1, HSW_PS_SAMPLE_MASK); /* 1 sample for now */
165
166 dw4 |= (brw->max_wm_threads - 1) << max_threads_shift;
167
168 /* CACHE_NEW_WM_PROG */
169 if (brw->wm.prog_data->nr_params > 0)
170 dw4 |= GEN7_PS_PUSH_CONSTANT_ENABLE;
171
172 /* CACHE_NEW_WM_PROG | _NEW_COLOR
173 *
174 * The hardware wedges if you have this bit set but don't turn on any dual
175 * source blend factors.
176 */
177 if (brw->wm.prog_data->dual_src_blend &&
178 (ctx->Color.BlendEnabled & 1) &&
179 ctx->Color.Blend[0]._UsesDualSrc) {
180 dw4 |= GEN7_PS_DUAL_SOURCE_BLEND_ENABLE;
181 }
182
183 /* CACHE_NEW_WM_PROG */
184 if (brw->wm.prog_data->num_varying_inputs != 0)
185 dw4 |= GEN7_PS_ATTRIBUTE_ENABLE;
186
187 dw4 |= GEN7_PS_8_DISPATCH_ENABLE;
188 if (brw->wm.prog_data->prog_offset_16)
189 dw4 |= GEN7_PS_16_DISPATCH_ENABLE;
190
191 dw5 |= (brw->wm.prog_data->first_curbe_grf <<
192 GEN7_PS_DISPATCH_START_GRF_SHIFT_0);
193 dw5 |= (brw->wm.prog_data->first_curbe_grf_16 <<
194 GEN7_PS_DISPATCH_START_GRF_SHIFT_2);
195
196 BEGIN_BATCH(8);
197 OUT_BATCH(_3DSTATE_PS << 16 | (8 - 2));
198 OUT_BATCH(brw->wm.base.prog_offset);
199 OUT_BATCH(dw2);
200 if (brw->wm.prog_data->total_scratch) {
201 OUT_RELOC(brw->wm.base.scratch_bo,
202 I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
203 ffs(brw->wm.prog_data->total_scratch) - 11);
204 } else {
205 OUT_BATCH(0);
206 }
207 OUT_BATCH(dw4);
208 OUT_BATCH(dw5);
209 OUT_BATCH(0); /* kernel 1 pointer */
210 OUT_BATCH(brw->wm.base.prog_offset + brw->wm.prog_data->prog_offset_16);
211 ADVANCE_BATCH();
212 }
213
214 const struct brw_tracked_state gen7_ps_state = {
215 .dirty = {
216 .mesa = (_NEW_PROGRAM_CONSTANTS |
217 _NEW_COLOR),
218 .brw = (BRW_NEW_FRAGMENT_PROGRAM |
219 BRW_NEW_PS_BINDING_TABLE |
220 BRW_NEW_BATCH |
221 BRW_NEW_PUSH_CONSTANT_ALLOCATION),
222 .cache = (CACHE_NEW_SAMPLER |
223 CACHE_NEW_WM_PROG)
224 },
225 .emit = upload_ps_state,
226 };