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
|
/*
* Copyright 2010 Red Hat 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
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* 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 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 NONINFRINGEMENT. IN NO EVENT SHALL
* THE COPYRIGHT HOLDER(S) OR AUTHOR(S) 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: Ben Skeggs
*/
#include <drm/drmP.h>
#include "nouveau_drv.h"
#include "nouveau_vm.h"
void
nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde,
struct nouveau_gpuobj *pgt[2])
{
u64 phys = 0xdeadcafe00000000ULL;
u32 coverage = 0;
if (pgt[0]) {
phys = 0x00000003 | pgt[0]->vinst; /* present, 4KiB pages */
coverage = (pgt[0]->size >> 3) << 12;
} else
if (pgt[1]) {
phys = 0x00000001 | pgt[1]->vinst; /* present */
coverage = (pgt[1]->size >> 3) << 16;
}
if (phys & 1) {
if (coverage <= 32 * 1024 * 1024)
phys |= 0x60;
else if (coverage <= 64 * 1024 * 1024)
phys |= 0x40;
else if (coverage <= 128 * 1024 * 1024)
phys |= 0x20;
}
nv_wo32(pgd, (pde * 8) + 0, lower_32_bits(phys));
nv_wo32(pgd, (pde * 8) + 4, upper_32_bits(phys));
}
static inline u64
vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target)
{
phys |= 1; /* present */
phys |= (u64)memtype << 40;
phys |= target << 4;
if (vma->access & NV_MEM_ACCESS_SYS)
phys |= (1 << 6);
if (!(vma->access & NV_MEM_ACCESS_WO))
phys |= (1 << 3);
return phys;
}
void
nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
struct nouveau_mem *mem, u32 pte, u32 cnt, u64 phys, u64 delta)
{
struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private;
u32 comp = (mem->memtype & 0x180) >> 7;
u32 block, target;
int i;
/* IGPs don't have real VRAM, re-target to stolen system memory */
target = 0;
if (dev_priv->vram_sys_base) {
phys += dev_priv->vram_sys_base;
target = 3;
}
phys = vm_addr(vma, phys, mem->memtype, target);
pte <<= 3;
cnt <<= 3;
while (cnt) {
u32 offset_h = upper_32_bits(phys);
u32 offset_l = lower_32_bits(phys);
for (i = 7; i >= 0; i--) {
block = 1 << (i + 3);
if (cnt >= block && !(pte & (block - 1)))
break;
}
offset_l |= (i << 7);
phys += block << (vma->node->type - 3);
cnt -= block;
if (comp) {
u32 tag = mem->tag->start + ((delta >> 16) * comp);
offset_h |= (tag << 17);
delta += block << (vma->node->type - 3);
}
while (block) {
nv_wo32(pgt, pte + 0, offset_l);
nv_wo32(pgt, pte + 4, offset_h);
pte += 8;
block -= 8;
}
}
}
void
nv50_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list)
{
u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 3 : 2;
pte <<= 3;
while (cnt--) {
u64 phys = vm_addr(vma, (u64)*list++, mem->memtype, target);
nv_wo32(pgt, pte + 0, lower_32_bits(phys));
nv_wo32(pgt, pte + 4, upper_32_bits(phys));
pte += 8;
}
}
void
nv50_vm_unmap(struct nouveau_gpuobj *pgt, u32 pte, u32 cnt)
{
pte <<= 3;
while (cnt--) {
nv_wo32(pgt, pte + 0, 0x00000000);
nv_wo32(pgt, pte + 4, 0x00000000);
pte += 8;
}
}
void
nv50_vm_flush(struct nouveau_vm *vm)
{
struct drm_nouveau_private *dev_priv = vm->dev->dev_private;
struct nouveau_instmem_engine *pinstmem = &dev_priv->engine.instmem;
int i;
pinstmem->flush(vm->dev);
/* BAR */
if (vm == dev_priv->bar1_vm || vm == dev_priv->bar3_vm) {
nv50_vm_flush_engine(vm->dev, 6);
return;
}
for (i = 0; i < NVOBJ_ENGINE_NR; i++) {
if (atomic_read(&vm->engref[i]))
dev_priv->eng[i]->tlb_flush(vm->dev, i);
}
}
void
nv50_vm_flush_engine(struct drm_device *dev, int engine)
{
struct drm_nouveau_private *dev_priv = dev->dev_private;
unsigned long flags;
spin_lock_irqsave(&dev_priv->vm_lock, flags);
nv_wr32(dev, 0x100c80, (engine << 16) | 1);
if (!nv_wait(dev, 0x100c80, 0x00000001, 0x00000000))
NV_ERROR(dev, "vm flush timeout: engine %d\n", engine);
spin_unlock_irqrestore(&dev_priv->vm_lock, flags);
}
|
