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
|
#include "pipe/p_context.h"
#include "pipe/p_state.h"
#include "pipe/p_util.h"
#include "nv40_context.h"
#include "nv40_dma.h"
#include "nv40_state.h"
#include "nvgl_pipe.h"
boolean
nv40_draw_arrays(struct pipe_context *pipe, unsigned mode, unsigned start,
unsigned count)
{
struct nv40_context *nv40 = (struct nv40_context *)pipe;
unsigned nr;
if (nv40->dirty)
nv40_emit_hw_state(nv40);
BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
OUT_RING (nvgl_primitive(mode));
nr = (count & 0xff);
if (nr) {
BEGIN_RING(curie, NV40TCL_VB_VERTEX_BATCH, 1);
OUT_RING (((nr - 1) << 24) | start);
start += nr;
}
/*XXX: large arrays (nr>2047) will blow up */
nr = count >> 8;
if (nr) {
assert (nr <= 2047);
BEGIN_RING_NI(curie, NV40TCL_VB_VERTEX_BATCH, nr);
while (nr--) {
OUT_RING(((0x100 - 1) << 24) | start);
start += 0x100;
}
}
BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
OUT_RING (0);
pipe->flush(pipe, PIPE_FLUSH_WAIT);
return TRUE;
}
static INLINE void
nv40_draw_elements_u08(struct nv40_context *nv40, void *ib,
unsigned start, unsigned count)
{
uint8_t *elts = (uint8_t *)ib + start;
int push, i;
if (count & 1) {
BEGIN_RING(curie, NV40TCL_VB_ELEMENT_U32, 1);
OUT_RING (elts[0]);
elts++; count--;
}
while (count) {
push = MIN2(count, 2046);
BEGIN_RING_NI(curie, NV40TCL_VB_ELEMENT_U16, push);
for (i = 0; i < push; i+=2)
OUT_RING((elts[i+1] << 16) | elts[i]);
count -= push;
elts += push;
}
}
static INLINE void
nv40_draw_elements_u16(struct nv40_context *nv40, void *ib,
unsigned start, unsigned count)
{
uint16_t *elts = (uint16_t *)ib + start;
int push, i;
if (count & 1) {
BEGIN_RING(curie, NV40TCL_VB_ELEMENT_U32, 1);
OUT_RING (elts[0]);
elts++; count--;
}
while (count) {
push = MIN2(count, 2046);
BEGIN_RING_NI(curie, NV40TCL_VB_ELEMENT_U16, push);
for (i = 0; i < push; i+=2)
OUT_RING((elts[i+1] << 16) | elts[i]);
count -= push;
elts += push;
}
}
static INLINE void
nv40_draw_elements_u32(struct nv40_context *nv40, void *ib,
unsigned start, unsigned count)
{
uint32_t *elts = (uint32_t *)ib + start;
int push;
while (count) {
push = MIN2(count, 2047);
BEGIN_RING_NI(curie, NV40TCL_VB_ELEMENT_U32, push);
OUT_RINGp (elts, push);
count -= push;
elts += push;
}
}
boolean
nv40_draw_elements(struct pipe_context *pipe,
struct pipe_buffer_handle *indexBuffer, unsigned indexSize,
unsigned mode, unsigned start, unsigned count)
{
struct nv40_context *nv40 = (struct nv40_context *)pipe;
void *ib;
if (nv40->dirty)
nv40_emit_hw_state(nv40);
ib = pipe->winsys->buffer_map(pipe->winsys, indexBuffer,
PIPE_BUFFER_FLAG_READ);
if (!ib) {
NOUVEAU_ERR("Couldn't map index buffer!!\n");
return FALSE;
}
BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
OUT_RING (nvgl_primitive(mode));
switch (indexSize) {
case 1:
nv40_draw_elements_u08(nv40, ib, start, count);
break;
case 2:
nv40_draw_elements_u16(nv40, ib, start, count);
break;
case 4:
nv40_draw_elements_u32(nv40, ib, start, count);
break;
default:
NOUVEAU_ERR("unsupported elt size %d\n", indexSize);
break;
}
BEGIN_RING(curie, NV40TCL_BEGIN_END, 1);
OUT_RING (0);
pipe->winsys->buffer_unmap(pipe->winsys, ib);
pipe->flush(pipe, PIPE_FLUSH_WAIT);
return TRUE;
}
static INLINE int
nv40_vbo_format_to_ncomp(uint format)
{
switch (format) {
case PIPE_FORMAT_R32G32B32A32_FLOAT: return 4;
case PIPE_FORMAT_R32G32B32_FLOAT: return 3;
case PIPE_FORMAT_R32G32_FLOAT: return 2;
case PIPE_FORMAT_R32_FLOAT: return 1;
default:
NOUVEAU_ERR("AII, unknown vbo format %d\n", format);
return 1;
}
}
void
nv40_vbo_arrays_update(struct nv40_context *nv40)
{
struct nouveau_winsys *nvws = nv40->nvws;
struct nv40_vertex_program *vp = nv40->vertprog.active;
uint32_t inputs, vtxfmt[16];
int hw, num_hw;
inputs = vp->ir;
for (hw = 0; hw < 16 && inputs; hw++) {
if (inputs & (1 << hw)) {
num_hw = hw;
inputs &= ~(1 << hw);
}
}
num_hw++;
inputs = vp->ir;
BEGIN_RING(curie, NV40TCL_VTXBUF_ADDRESS(0), num_hw);
for (hw = 0; hw < num_hw; hw++) {
struct pipe_vertex_element *ve;
struct pipe_vertex_buffer *vb;
if (!(inputs & (1 << hw))) {
OUT_RING(0);
vtxfmt[hw] = NV40TCL_VTXFMT_TYPE_FLOAT;
continue;
}
ve = &nv40->vtxelt[hw];
vb = &nv40->vtxbuf[ve->vertex_buffer_index];
OUT_RELOC(vb->buffer, vb->buffer_offset + ve->src_offset,
NOUVEAU_BO_GART | NOUVEAU_BO_VRAM | NOUVEAU_BO_LOW |
NOUVEAU_BO_OR | NOUVEAU_BO_RD, 0,
NV40TCL_VTXBUF_ADDRESS_DMA1);
vtxfmt[hw] = ((vb->pitch << NV40TCL_VTXFMT_STRIDE_SHIFT) |
(nv40_vbo_format_to_ncomp(ve->src_format) <<
NV40TCL_VTXFMT_SIZE_SHIFT) |
NV40TCL_VTXFMT_TYPE_FLOAT);
}
BEGIN_RING(curie, 0x1710, 1);
OUT_RING (0); /* vtx cache flush */
BEGIN_RING(curie, NV40TCL_VTXFMT(0), num_hw);
OUT_RINGp (vtxfmt, num_hw);
}
|
