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
|
/*
* Copyright (c) 2017, NVIDIA CORPORATION. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
* version 2, as published by the Free Software Foundation.
*
* This program is distributed in the hope it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <nvgpu/pramin.h>
#include <nvgpu/page_allocator.h>
#include "gk20a/gk20a.h"
/*
* Flip this to force all gk20a_mem* accesses via PRAMIN from the start of the
* boot, even for buffers that would work via cpu_va. In runtime, the flag is
* in debugfs, called "force_pramin".
*/
#define GK20A_FORCE_PRAMIN_DEFAULT false
void pramin_access_batch_rd_n(struct gk20a *g, u32 start, u32 words, u32 **arg)
{
u32 r = start, *dest_u32 = *arg;
if (!g->regs) {
__gk20a_warn_on_no_regs();
return;
}
while (words--) {
*dest_u32++ = gk20a_readl(g, r);
r += sizeof(u32);
}
*arg = dest_u32;
}
void pramin_access_batch_wr_n(struct gk20a *g, u32 start, u32 words, u32 **arg)
{
u32 r = start, *src_u32 = *arg;
if (!g->regs) {
__gk20a_warn_on_no_regs();
return;
}
while (words--) {
writel_relaxed(*src_u32++, g->regs + r);
r += sizeof(u32);
}
*arg = src_u32;
}
void pramin_access_batch_set(struct gk20a *g, u32 start, u32 words, u32 **arg)
{
u32 r = start, repeat = **arg;
if (!g->regs) {
__gk20a_warn_on_no_regs();
return;
}
while (words--) {
writel_relaxed(repeat, g->regs + r);
r += sizeof(u32);
}
}
/*
* The PRAMIN range is 1 MB, must change base addr if a buffer crosses that.
* This same loop is used for read/write/memset. Offset and size in bytes.
* One call to "loop" is done per range, with "arg" supplied.
*/
void nvgpu_pramin_access_batched(struct gk20a *g, struct mem_desc *mem,
u32 offset, u32 size, pramin_access_batch_fn loop, u32 **arg)
{
struct nvgpu_page_alloc *alloc = NULL;
struct page_alloc_chunk *chunk = NULL;
u32 byteoff, start_reg, until_end, n;
alloc = get_vidmem_page_alloc(mem->sgt->sgl);
nvgpu_list_for_each_entry(chunk, &alloc->alloc_chunks,
page_alloc_chunk, list_entry) {
if (offset >= chunk->length)
offset -= chunk->length;
else
break;
}
offset /= sizeof(u32);
while (size) {
byteoff = g->ops.pramin.enter(g, mem, chunk, offset);
start_reg = g->ops.pramin.data032_r(byteoff / sizeof(u32));
until_end = SZ_1M - (byteoff & (SZ_1M - 1));
n = min3(size, until_end, (u32)(chunk->length - offset));
loop(g, start_reg, n / sizeof(u32), arg);
/* read back to synchronize accesses */
gk20a_readl(g, start_reg);
g->ops.pramin.exit(g, mem, chunk);
size -= n;
if (n == (chunk->length - offset)) {
chunk = nvgpu_list_next_entry(chunk, page_alloc_chunk,
list_entry);
offset = 0;
} else {
offset += n / sizeof(u32);
}
}
}
void nvgpu_init_pramin(struct mm_gk20a *mm)
{
mm->pramin_window = 0;
nvgpu_spinlock_init(&mm->pramin_window_lock);
mm->force_pramin = GK20A_FORCE_PRAMIN_DEFAULT;
}
|
