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
|
/*
* Copyright 2013 Advanced Micro Devices, Inc.
*
* 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
* on the rights to use, copy, modify, merge, publish, distribute, sub
* license, 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 NON-INFRINGEMENT. IN NO EVENT SHALL
* THE AUTHOR(S) AND/OR THEIR SUPPLIERS 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.
*
* Authors: Marek Olšák <maraeo@gmail.com>
*/
/**
* This file contains helpers for writing commands to commands streams.
*/
#ifndef R600_CS_H
#define R600_CS_H
#include "r600_pipe_common.h"
#include "r600d_common.h"
/**
* Return true if there is enough memory in VRAM and GTT for the buffers
* added so far.
*
* \param vram VRAM memory size not added to the buffer list yet
* \param gtt GTT memory size not added to the buffer list yet
*/
static inline bool
radeon_cs_memory_below_limit(struct r600_common_screen *screen,
struct radeon_winsys_cs *cs,
uint64_t vram, uint64_t gtt)
{
vram += cs->used_vram;
gtt += cs->used_gart;
/* Anything that goes above the VRAM size should go to GTT. */
if (vram > screen->info.vram_size)
gtt += vram - screen->info.vram_size;
/* Now we just need to check if we have enough GTT. */
return gtt < screen->info.gart_size * 0.7;
}
/**
* Add a buffer to the buffer list for the given command stream (CS).
*
* All buffers used by a CS must be added to the list. This tells the kernel
* driver which buffers are used by GPU commands. Other buffers can
* be swapped out (not accessible) during execution.
*
* The buffer list becomes empty after every context flush and must be
* rebuilt.
*/
static inline unsigned radeon_add_to_buffer_list(struct r600_common_context *rctx,
struct r600_ring *ring,
struct r600_resource *rbo,
enum radeon_bo_usage usage,
enum radeon_bo_priority priority)
{
assert(usage);
return rctx->ws->cs_add_buffer(ring->cs, rbo->buf, usage,
rbo->domains, priority) * 4;
}
/**
* Same as above, but also checks memory usage and flushes the context
* accordingly.
*
* When this SHOULD NOT be used:
*
* - if r600_context_add_resource_size has been called for the buffer
* followed by *_need_cs_space for checking the memory usage
*
* - if r600_need_dma_space has been called for the buffer
*
* - when emitting state packets and draw packets (because preceding packets
* can't be re-emitted at that point)
*
* - if shader resource "enabled_mask" is not up-to-date or there is
* a different constraint disallowing a context flush
*/
static inline unsigned
radeon_add_to_buffer_list_check_mem(struct r600_common_context *rctx,
struct r600_ring *ring,
struct r600_resource *rbo,
enum radeon_bo_usage usage,
enum radeon_bo_priority priority,
bool check_mem)
{
if (check_mem &&
!radeon_cs_memory_below_limit(rctx->screen, ring->cs,
rctx->vram + rbo->vram_usage,
rctx->gtt + rbo->gart_usage))
ring->flush(rctx, RADEON_FLUSH_ASYNC, NULL);
return radeon_add_to_buffer_list(rctx, ring, rbo, usage, priority);
}
static inline void r600_emit_reloc(struct r600_common_context *rctx,
struct r600_ring *ring, struct r600_resource *rbo,
enum radeon_bo_usage usage,
enum radeon_bo_priority priority)
{
struct radeon_winsys_cs *cs = ring->cs;
bool has_vm = ((struct r600_common_screen*)rctx->b.screen)->info.has_virtual_memory;
unsigned reloc = radeon_add_to_buffer_list(rctx, ring, rbo, usage, priority);
if (!has_vm) {
radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
radeon_emit(cs, reloc);
}
}
static inline void radeon_set_config_reg_seq(struct radeon_winsys_cs *cs, unsigned reg, unsigned num)
{
assert(reg < R600_CONTEXT_REG_OFFSET);
assert(cs->current.cdw + 2 + num <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_CONFIG_REG, num, 0));
radeon_emit(cs, (reg - R600_CONFIG_REG_OFFSET) >> 2);
}
static inline void radeon_set_config_reg(struct radeon_winsys_cs *cs, unsigned reg, unsigned value)
{
radeon_set_config_reg_seq(cs, reg, 1);
radeon_emit(cs, value);
}
static inline void radeon_set_context_reg_seq(struct radeon_winsys_cs *cs, unsigned reg, unsigned num)
{
assert(reg >= R600_CONTEXT_REG_OFFSET);
assert(cs->current.cdw + 2 + num <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_CONTEXT_REG, num, 0));
radeon_emit(cs, (reg - R600_CONTEXT_REG_OFFSET) >> 2);
}
static inline void radeon_set_context_reg(struct radeon_winsys_cs *cs, unsigned reg, unsigned value)
{
radeon_set_context_reg_seq(cs, reg, 1);
radeon_emit(cs, value);
}
static inline void radeon_set_context_reg_idx(struct radeon_winsys_cs *cs,
unsigned reg, unsigned idx,
unsigned value)
{
assert(reg >= R600_CONTEXT_REG_OFFSET);
assert(cs->current.cdw + 3 <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_CONTEXT_REG, 1, 0));
radeon_emit(cs, (reg - R600_CONTEXT_REG_OFFSET) >> 2 | (idx << 28));
radeon_emit(cs, value);
}
static inline void radeon_set_sh_reg_seq(struct radeon_winsys_cs *cs, unsigned reg, unsigned num)
{
assert(reg >= SI_SH_REG_OFFSET && reg < SI_SH_REG_END);
assert(cs->current.cdw + 2 + num <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_SH_REG, num, 0));
radeon_emit(cs, (reg - SI_SH_REG_OFFSET) >> 2);
}
static inline void radeon_set_sh_reg(struct radeon_winsys_cs *cs, unsigned reg, unsigned value)
{
radeon_set_sh_reg_seq(cs, reg, 1);
radeon_emit(cs, value);
}
static inline void radeon_set_uconfig_reg_seq(struct radeon_winsys_cs *cs, unsigned reg, unsigned num)
{
assert(reg >= CIK_UCONFIG_REG_OFFSET && reg < CIK_UCONFIG_REG_END);
assert(cs->current.cdw + 2 + num <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_UCONFIG_REG, num, 0));
radeon_emit(cs, (reg - CIK_UCONFIG_REG_OFFSET) >> 2);
}
static inline void radeon_set_uconfig_reg(struct radeon_winsys_cs *cs, unsigned reg, unsigned value)
{
radeon_set_uconfig_reg_seq(cs, reg, 1);
radeon_emit(cs, value);
}
static inline void radeon_set_uconfig_reg_idx(struct radeon_winsys_cs *cs,
unsigned reg, unsigned idx,
unsigned value)
{
assert(reg >= CIK_UCONFIG_REG_OFFSET && reg < CIK_UCONFIG_REG_END);
assert(cs->current.cdw + 3 <= cs->current.max_dw);
radeon_emit(cs, PKT3(PKT3_SET_UCONFIG_REG, 1, 0));
radeon_emit(cs, (reg - CIK_UCONFIG_REG_OFFSET) >> 2 | (idx << 28));
radeon_emit(cs, value);
}
#endif
|
