2 * Copyright © 2011 Intel Corporation
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:
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
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
27 * This file computes the "VUE map" for a (non-fragment) shader stage, which
28 * describes the layout of its output varyings. The VUE map is used to match
29 * outputs from one stage with the inputs of the next.
31 * Largely, varyings can be placed however we like - producers/consumers simply
32 * have to agree on the layout. However, there is also a "VUE Header" that
33 * prescribes a fixed-layout for items that interact with fixed function
34 * hardware, such as the clipper and rasterizer.
37 * Paul Berry <stereotype441@gmail.com>
38 * Chris Forbes <chrisf@ijw.co.nz>
39 * Eric Anholt <eric@anholt.net>
43 #include "main/compiler.h"
44 #include "brw_context.h"
47 assign_vue_slot(struct brw_vue_map
*vue_map
, int varying
)
49 /* Make sure this varying hasn't been assigned a slot already */
50 assert (vue_map
->varying_to_slot
[varying
] == -1);
52 vue_map
->varying_to_slot
[varying
] = vue_map
->num_slots
;
53 vue_map
->slot_to_varying
[vue_map
->num_slots
++] = varying
;
57 * Compute the VUE map for a shader stage.
60 brw_compute_vue_map(const struct brw_device_info
*devinfo
,
61 struct brw_vue_map
*vue_map
,
62 GLbitfield64 slots_valid
)
64 vue_map
->slots_valid
= slots_valid
;
67 /* gl_Layer and gl_ViewportIndex don't get their own varying slots -- they
68 * are stored in the first VUE slot (VARYING_SLOT_PSIZ).
70 slots_valid
&= ~(VARYING_BIT_LAYER
| VARYING_BIT_VIEWPORT
);
72 /* Make sure that the values we store in vue_map->varying_to_slot and
73 * vue_map->slot_to_varying won't overflow the signed chars that are used
74 * to store them. Note that since vue_map->slot_to_varying sometimes holds
75 * values equal to BRW_VARYING_SLOT_COUNT, we need to ensure that
76 * BRW_VARYING_SLOT_COUNT is <= 127, not 128.
78 STATIC_ASSERT(BRW_VARYING_SLOT_COUNT
<= 127);
80 vue_map
->num_slots
= 0;
81 for (i
= 0; i
< BRW_VARYING_SLOT_COUNT
; ++i
) {
82 vue_map
->varying_to_slot
[i
] = -1;
83 vue_map
->slot_to_varying
[i
] = BRW_VARYING_SLOT_COUNT
;
86 /* VUE header: format depends on chip generation and whether clipping is
89 * See the Sandybridge PRM, Volume 2 Part 1, section 1.5.1 (page 30),
90 * "Vertex URB Entry (VUE) Formats" which describes the VUE header layout.
92 if (devinfo
->gen
< 6) {
93 /* There are 8 dwords in VUE header pre-Ironlake:
94 * dword 0-3 is indices, point width, clip flags.
95 * dword 4-7 is ndc position
96 * dword 8-11 is the first vertex data.
98 * On Ironlake the VUE header is nominally 20 dwords, but the hardware
99 * will accept the same header layout as Gen4 [and should be a bit faster]
101 assign_vue_slot(vue_map
, VARYING_SLOT_PSIZ
);
102 assign_vue_slot(vue_map
, BRW_VARYING_SLOT_NDC
);
103 assign_vue_slot(vue_map
, VARYING_SLOT_POS
);
105 /* There are 8 or 16 DWs (D0-D15) in VUE header on Sandybridge:
106 * dword 0-3 of the header is indices, point width, clip flags.
107 * dword 4-7 is the 4D space position
108 * dword 8-15 of the vertex header is the user clip distance if
110 * dword 8-11 or 16-19 is the first vertex element data we fill.
112 assign_vue_slot(vue_map
, VARYING_SLOT_PSIZ
);
113 assign_vue_slot(vue_map
, VARYING_SLOT_POS
);
114 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST0
))
115 assign_vue_slot(vue_map
, VARYING_SLOT_CLIP_DIST0
);
116 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_CLIP_DIST1
))
117 assign_vue_slot(vue_map
, VARYING_SLOT_CLIP_DIST1
);
119 /* front and back colors need to be consecutive so that we can use
120 * ATTRIBUTE_SWIZZLE_INPUTATTR_FACING to swizzle them when doing
123 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_COL0
))
124 assign_vue_slot(vue_map
, VARYING_SLOT_COL0
);
125 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_BFC0
))
126 assign_vue_slot(vue_map
, VARYING_SLOT_BFC0
);
127 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_COL1
))
128 assign_vue_slot(vue_map
, VARYING_SLOT_COL1
);
129 if (slots_valid
& BITFIELD64_BIT(VARYING_SLOT_BFC1
))
130 assign_vue_slot(vue_map
, VARYING_SLOT_BFC1
);
133 /* The hardware doesn't care about the rest of the vertex outputs, so just
134 * assign them contiguously. Don't reassign outputs that already have a
137 * We generally don't need to assign a slot for VARYING_SLOT_CLIP_VERTEX,
138 * since it's encoded as the clip distances by emit_clip_distances().
139 * However, it may be output by transform feedback, and we'd rather not
140 * recompute state when TF changes, so we just always include it.
142 for (int i
= 0; i
< VARYING_SLOT_MAX
; ++i
) {
143 if ((slots_valid
& BITFIELD64_BIT(i
)) &&
144 vue_map
->varying_to_slot
[i
] == -1) {
145 assign_vue_slot(vue_map
, i
);