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
|
// SPDX-License-Identifier: GPL-2.0
/*
* In-kernel vector facility support functions
*
* Copyright IBM Corp. 2015
* Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
*/
#include <linux/kernel.h>
#include <linux/cpu.h>
#include <linux/sched.h>
#include <asm/fpu.h>
void __kernel_fpu_begin(struct kernel_fpu *state, int flags)
{
__vector128 *vxrs = state->vxrs;
int mask;
/*
* Limit the save to the FPU/vector registers already
* in use by the previous context.
*/
flags &= state->hdr.mask;
if (flags & KERNEL_FPC)
fpu_stfpc(&state->hdr.fpc);
if (!cpu_has_vx()) {
if (flags & KERNEL_VXR_LOW)
save_fp_regs_vx(vxrs);
return;
}
mask = flags & KERNEL_VXR;
if (mask == KERNEL_VXR) {
vxrs += fpu_vstm(0, 15, vxrs);
vxrs += fpu_vstm(16, 31, vxrs);
return;
}
if (mask == KERNEL_VXR_MID) {
vxrs += fpu_vstm(8, 23, vxrs);
return;
}
mask = flags & KERNEL_VXR_LOW;
if (mask) {
if (mask == KERNEL_VXR_LOW)
vxrs += fpu_vstm(0, 15, vxrs);
else if (mask == KERNEL_VXR_V0V7)
vxrs += fpu_vstm(0, 7, vxrs);
else
vxrs += fpu_vstm(8, 15, vxrs);
}
mask = flags & KERNEL_VXR_HIGH;
if (mask) {
if (mask == KERNEL_VXR_HIGH)
vxrs += fpu_vstm(16, 31, vxrs);
else if (mask == KERNEL_VXR_V16V23)
vxrs += fpu_vstm(16, 23, vxrs);
else
vxrs += fpu_vstm(24, 31, vxrs);
}
}
EXPORT_SYMBOL(__kernel_fpu_begin);
void __kernel_fpu_end(struct kernel_fpu *state, int flags)
{
__vector128 *vxrs = state->vxrs;
int mask;
/*
* Limit the restore to the FPU/vector registers of the
* previous context that have been overwritten by the
* current context.
*/
flags &= state->hdr.mask;
if (flags & KERNEL_FPC)
fpu_lfpc(&state->hdr.fpc);
if (!cpu_has_vx()) {
if (flags & KERNEL_VXR_LOW)
load_fp_regs_vx(vxrs);
return;
}
mask = flags & KERNEL_VXR;
if (mask == KERNEL_VXR) {
vxrs += fpu_vlm(0, 15, vxrs);
vxrs += fpu_vlm(16, 31, vxrs);
return;
}
if (mask == KERNEL_VXR_MID) {
vxrs += fpu_vlm(8, 23, vxrs);
return;
}
mask = flags & KERNEL_VXR_LOW;
if (mask) {
if (mask == KERNEL_VXR_LOW)
vxrs += fpu_vlm(0, 15, vxrs);
else if (mask == KERNEL_VXR_V0V7)
vxrs += fpu_vlm(0, 7, vxrs);
else
vxrs += fpu_vlm(8, 15, vxrs);
}
mask = flags & KERNEL_VXR_HIGH;
if (mask) {
if (mask == KERNEL_VXR_HIGH)
vxrs += fpu_vlm(16, 31, vxrs);
else if (mask == KERNEL_VXR_V16V23)
vxrs += fpu_vlm(16, 23, vxrs);
else
vxrs += fpu_vlm(24, 31, vxrs);
}
}
EXPORT_SYMBOL(__kernel_fpu_end);
void load_fpu_state(struct fpu *state, int flags)
{
__vector128 *vxrs = &state->vxrs[0];
int mask;
if (flags & KERNEL_FPC)
fpu_lfpc(&state->fpc);
if (!cpu_has_vx()) {
if (flags & KERNEL_VXR_V0V7)
load_fp_regs_vx(state->vxrs);
return;
}
mask = flags & KERNEL_VXR;
if (mask == KERNEL_VXR) {
fpu_vlm(0, 15, &vxrs[0]);
fpu_vlm(16, 31, &vxrs[16]);
return;
}
if (mask == KERNEL_VXR_MID) {
fpu_vlm(8, 23, &vxrs[8]);
return;
}
mask = flags & KERNEL_VXR_LOW;
if (mask) {
if (mask == KERNEL_VXR_LOW)
fpu_vlm(0, 15, &vxrs[0]);
else if (mask == KERNEL_VXR_V0V7)
fpu_vlm(0, 7, &vxrs[0]);
else
fpu_vlm(8, 15, &vxrs[8]);
}
mask = flags & KERNEL_VXR_HIGH;
if (mask) {
if (mask == KERNEL_VXR_HIGH)
fpu_vlm(16, 31, &vxrs[16]);
else if (mask == KERNEL_VXR_V16V23)
fpu_vlm(16, 23, &vxrs[16]);
else
fpu_vlm(24, 31, &vxrs[24]);
}
}
void save_fpu_state(struct fpu *state, int flags)
{
__vector128 *vxrs = &state->vxrs[0];
int mask;
if (flags & KERNEL_FPC)
fpu_stfpc(&state->fpc);
if (!cpu_has_vx()) {
if (flags & KERNEL_VXR_LOW)
save_fp_regs_vx(state->vxrs);
return;
}
mask = flags & KERNEL_VXR;
if (mask == KERNEL_VXR) {
fpu_vstm(0, 15, &vxrs[0]);
fpu_vstm(16, 31, &vxrs[16]);
return;
}
if (mask == KERNEL_VXR_MID) {
fpu_vstm(8, 23, &vxrs[8]);
return;
}
mask = flags & KERNEL_VXR_LOW;
if (mask) {
if (mask == KERNEL_VXR_LOW)
fpu_vstm(0, 15, &vxrs[0]);
else if (mask == KERNEL_VXR_V0V7)
fpu_vstm(0, 7, &vxrs[0]);
else
fpu_vstm(8, 15, &vxrs[8]);
}
mask = flags & KERNEL_VXR_HIGH;
if (mask) {
if (mask == KERNEL_VXR_HIGH)
fpu_vstm(16, 31, &vxrs[16]);
else if (mask == KERNEL_VXR_V16V23)
fpu_vstm(16, 23, &vxrs[16]);
else
fpu_vstm(24, 31, &vxrs[24]);
}
}
EXPORT_SYMBOL(save_fpu_state);
|
