i915: Remove most of the code under gen >= 4 checks.
[mesa.git] / src / mesa / drivers / dri / i965 / gen7_urb.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 "main/macros.h"
25 #include "intel_batchbuffer.h"
26 #include "brw_context.h"
27 #include "brw_state.h"
28 #include "brw_defines.h"
29
30 /**
31 * The following diagram shows how we partition the URB:
32 *
33 * 8kB 8kB Rest of the URB space
34 * ____-____ ____-____ _________________-_________________
35 * / \ / \ / \
36 * +-------------------------------------------------------------+
37 * | VS Push | FS Push | VS |
38 * | Constants | Constants | Handles |
39 * +-------------------------------------------------------------+
40 *
41 * Notably, push constants must be stored at the beginning of the URB
42 * space, while entries can be stored anywhere. Ivybridge and Haswell
43 * GT1/GT2 have a maximum constant buffer size of 16kB, while Haswell GT3
44 * doubles this (32kB).
45 *
46 * Currently we split the constant buffer space evenly between VS and FS.
47 * This is probably not ideal, but simple.
48 *
49 * Ivybridge GT1 and Haswell GT1 have 128kB of URB space.
50 * Ivybridge GT2 and Haswell GT2 have 256kB of URB space.
51 * Haswell GT3 has 512kB of URB space.
52 *
53 * See "Volume 2a: 3D Pipeline," section 1.8, "Volume 1b: Configurations",
54 * and the documentation for 3DSTATE_PUSH_CONSTANT_ALLOC_xS.
55 */
56 void
57 gen7_allocate_push_constants(struct brw_context *brw)
58 {
59 struct intel_context *intel = &brw->intel;
60
61 unsigned size = 8;
62 if (intel->is_haswell && intel->gt == 3)
63 size = 16;
64
65 BEGIN_BATCH(2);
66 OUT_BATCH(_3DSTATE_PUSH_CONSTANT_ALLOC_VS << 16 | (2 - 2));
67 OUT_BATCH(size);
68 ADVANCE_BATCH();
69
70 BEGIN_BATCH(2);
71 OUT_BATCH(_3DSTATE_PUSH_CONSTANT_ALLOC_PS << 16 | (2 - 2));
72 OUT_BATCH(size | size << GEN7_PUSH_CONSTANT_BUFFER_OFFSET_SHIFT);
73 ADVANCE_BATCH();
74 }
75
76 static void
77 gen7_upload_urb(struct brw_context *brw)
78 {
79 struct intel_context *intel = &brw->intel;
80 const int push_size_kB = intel->is_haswell && intel->gt == 3 ? 32 : 16;
81
82 /* Total space for entries is URB size - 16kB for push constants */
83 int handle_region_size = (brw->urb.size - push_size_kB) * 1024; /* bytes */
84
85 /* CACHE_NEW_VS_PROG */
86 unsigned vs_size = MAX2(brw->vs.prog_data->base.urb_entry_size, 1);
87
88 int nr_vs_entries = handle_region_size / (vs_size * 64);
89 if (nr_vs_entries > brw->urb.max_vs_entries)
90 nr_vs_entries = brw->urb.max_vs_entries;
91
92 /* According to volume 2a, nr_vs_entries must be a multiple of 8. */
93 brw->urb.nr_vs_entries = ROUND_DOWN_TO(nr_vs_entries, 8);
94
95 /* URB Starting Addresses are specified in multiples of 8kB. */
96 brw->urb.vs_start = push_size_kB / 8; /* skip over push constants */
97
98 assert(brw->urb.nr_vs_entries % 8 == 0);
99 assert(brw->urb.nr_gs_entries % 8 == 0);
100 /* GS requirement */
101 assert(!brw->gs.prog_active);
102
103 gen7_emit_vs_workaround_flush(intel);
104 gen7_emit_urb_state(brw, brw->urb.nr_vs_entries, vs_size, brw->urb.vs_start);
105 }
106
107 void
108 gen7_emit_urb_state(struct brw_context *brw, GLuint nr_vs_entries,
109 GLuint vs_size, GLuint vs_start)
110 {
111 struct intel_context *intel = &brw->intel;
112
113 BEGIN_BATCH(2);
114 OUT_BATCH(_3DSTATE_URB_VS << 16 | (2 - 2));
115 OUT_BATCH(nr_vs_entries |
116 ((vs_size - 1) << GEN7_URB_ENTRY_SIZE_SHIFT) |
117 (vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
118 ADVANCE_BATCH();
119
120 /* Allocate the GS, HS, and DS zero space - we don't use them. */
121 BEGIN_BATCH(2);
122 OUT_BATCH(_3DSTATE_URB_GS << 16 | (2 - 2));
123 OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
124 (vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
125 ADVANCE_BATCH();
126
127 BEGIN_BATCH(2);
128 OUT_BATCH(_3DSTATE_URB_HS << 16 | (2 - 2));
129 OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
130 (vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
131 ADVANCE_BATCH();
132
133 BEGIN_BATCH(2);
134 OUT_BATCH(_3DSTATE_URB_DS << 16 | (2 - 2));
135 OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
136 (vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
137 ADVANCE_BATCH();
138 }
139
140 const struct brw_tracked_state gen7_urb = {
141 .dirty = {
142 .mesa = 0,
143 .brw = BRW_NEW_CONTEXT,
144 .cache = (CACHE_NEW_VS_PROG | CACHE_NEW_GS_PROG),
145 },
146 .emit = gen7_upload_urb,
147 };