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
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
|
/* linuxkm_memory.c
*
* Copyright (C) 2006-2022 wolfSSL Inc.
*
* This file is part of wolfSSL.
*
* wolfSSL is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* wolfSSL is distributed in the hope that 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, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
*/
/* included by wolfcrypt/src/memory.c */
#if defined(WOLFSSL_LINUXKM_SIMD_X86)
#ifdef LINUXKM_SIMD_IRQ
#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 16, 0)
static union fpregs_state **wolfcrypt_linuxkm_fpu_states = NULL;
#else
static struct fpstate **wolfcrypt_linuxkm_fpu_states = NULL;
#endif
#else
static unsigned int *wolfcrypt_linuxkm_fpu_states = NULL;
#endif
static WARN_UNUSED_RESULT inline int am_in_hard_interrupt_handler(void)
{
return (preempt_count() & (NMI_MASK | HARDIRQ_MASK)) != 0;
}
WARN_UNUSED_RESULT int allocate_wolfcrypt_linuxkm_fpu_states(void)
{
#ifdef LINUXKM_SIMD_IRQ
#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 16, 0)
wolfcrypt_linuxkm_fpu_states =
(union fpregs_state **)kzalloc(nr_cpu_ids
* sizeof(struct fpu_state *),
GFP_KERNEL);
#else
wolfcrypt_linuxkm_fpu_states =
(struct fpstate **)kzalloc(nr_cpu_ids
* sizeof(struct fpstate *),
GFP_KERNEL);
#endif
#else
wolfcrypt_linuxkm_fpu_states =
(unsigned int *)kzalloc(nr_cpu_ids * sizeof(unsigned int),
GFP_KERNEL);
#endif
if (! wolfcrypt_linuxkm_fpu_states) {
pr_err("warning, allocation of %lu bytes for "
"wolfcrypt_linuxkm_fpu_states failed.\n",
nr_cpu_ids * sizeof(struct fpu_state *));
return MEMORY_E;
}
#ifdef LINUXKM_SIMD_IRQ
{
typeof(nr_cpu_ids) i;
for (i=0; i<nr_cpu_ids; ++i) {
_Static_assert(sizeof(union fpregs_state) <= PAGE_SIZE,
"union fpregs_state is larger than expected.");
#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 16, 0)
wolfcrypt_linuxkm_fpu_states[i] =
(union fpregs_state *)kzalloc(PAGE_SIZE
/* sizeof(union fpregs_state) */,
GFP_KERNEL);
#else
wolfcrypt_linuxkm_fpu_states[i] =
(struct fpstate *)kzalloc(PAGE_SIZE
/* sizeof(struct fpstate) */,
GFP_KERNEL);
#endif
if (! wolfcrypt_linuxkm_fpu_states[i])
break;
/* double-check that the allocation is 64-byte-aligned as needed
* for xsave.
*/
if ((unsigned long)wolfcrypt_linuxkm_fpu_states[i] & 63UL) {
pr_err("warning, allocation for wolfcrypt_linuxkm_fpu_states "
"was not properly aligned (%px).\n",
wolfcrypt_linuxkm_fpu_states[i]);
kfree(wolfcrypt_linuxkm_fpu_states[i]);
wolfcrypt_linuxkm_fpu_states[i] = 0;
break;
}
}
if (i < nr_cpu_ids) {
pr_err("warning, only %u/%u allocations succeeded for "
"wolfcrypt_linuxkm_fpu_states.\n",
i, nr_cpu_ids);
return MEMORY_E;
}
}
#endif /* LINUXKM_SIMD_IRQ */
return 0;
}
void free_wolfcrypt_linuxkm_fpu_states(void)
{
if (wolfcrypt_linuxkm_fpu_states) {
#ifdef LINUXKM_SIMD_IRQ
typeof(nr_cpu_ids) i;
for (i=0; i<nr_cpu_ids; ++i) {
if (wolfcrypt_linuxkm_fpu_states[i])
kfree(wolfcrypt_linuxkm_fpu_states[i]);
}
#endif /* LINUXKM_SIMD_IRQ */
kfree(wolfcrypt_linuxkm_fpu_states);
wolfcrypt_linuxkm_fpu_states = 0;
}
}
WARN_UNUSED_RESULT int save_vector_registers_x86(void)
{
int processor_id;
preempt_disable();
processor_id = smp_processor_id();
{
static int _warned_on_null = -1;
if ((wolfcrypt_linuxkm_fpu_states == NULL)
#ifdef LINUXKM_SIMD_IRQ
|| (wolfcrypt_linuxkm_fpu_states[processor_id] == NULL)
#endif
)
{
preempt_enable();
if (_warned_on_null < processor_id) {
_warned_on_null = processor_id;
pr_err("save_vector_registers_x86 called for cpu id %d "
"with null context buffer.\n", processor_id);
}
return BAD_STATE_E;
}
}
if (! irq_fpu_usable()) {
#ifdef LINUXKM_SIMD_IRQ
if (am_in_hard_interrupt_handler()) {
/* allow for nested calls */
if (((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] != 0) {
if (((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] == 255) {
preempt_enable();
pr_err("save_vector_registers_x86 recursion register overflow for "
"cpu id %d.\n", processor_id);
return BAD_STATE_E;
} else {
++((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1];
return 0;
}
}
/* note, fpregs_lock() is not needed here, because
* interrupts/preemptions are already disabled here.
*/
{
/* save_fpregs_to_fpstate() only accesses fpu->state, which
* has stringent alignment requirements (64 byte cache
* line), but takes a pointer to the parent struct. work
* around this.
*/
#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 14, 0)
struct fpu *fake_fpu_pointer =
(struct fpu *)(((char *)wolfcrypt_linuxkm_fpu_states[processor_id])
- offsetof(struct fpu, state));
copy_fpregs_to_fpstate(fake_fpu_pointer);
#elif LINUX_VERSION_CODE < KERNEL_VERSION(5, 16, 0)
struct fpu *fake_fpu_pointer =
(struct fpu *)(((char *)wolfcrypt_linuxkm_fpu_states[processor_id])
- offsetof(struct fpu, state));
save_fpregs_to_fpstate(fake_fpu_pointer);
#else
struct fpu *fake_fpu_pointer =
(struct fpu *)(((char *)wolfcrypt_linuxkm_fpu_states[processor_id])
- offsetof(struct fpu, fpstate));
save_fpregs_to_fpstate(fake_fpu_pointer);
#endif
}
/* mark the slot as used. */
((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] = 1;
/* note, not preempt_enable()ing, mirroring kernel_fpu_begin()
* semantics, even though routine will have been entered already
* non-preemptable.
*/
return 0;
} else
#endif /* LINUXKM_SIMD_IRQ */
{
preempt_enable();
return BAD_STATE_E;
}
} else {
/* allow for nested calls */
#ifdef LINUXKM_SIMD_IRQ
if (((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] != 0) {
if (((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] == 255) {
preempt_enable();
pr_err("save_vector_registers_x86 recursion register overflow for "
"cpu id %d.\n", processor_id);
return BAD_STATE_E;
} else {
++((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1];
return 0;
}
}
kernel_fpu_begin();
preempt_enable(); /* kernel_fpu_begin() does its own
* preempt_disable(). decrement ours.
*/
((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] = 1;
#else /* !LINUXKM_SIMD_IRQ */
if (wolfcrypt_linuxkm_fpu_states[processor_id] != 0) {
if (wolfcrypt_linuxkm_fpu_states[processor_id] == ~0U) {
preempt_enable();
pr_err("save_vector_registers_x86 recursion register overflow for "
"cpu id %d.\n", processor_id);
return BAD_STATE_E;
} else {
++wolfcrypt_linuxkm_fpu_states[processor_id];
return 0;
}
}
kernel_fpu_begin();
preempt_enable(); /* kernel_fpu_begin() does its own
* preempt_disable(). decrement ours.
*/
wolfcrypt_linuxkm_fpu_states[processor_id] = 1;
#endif /* !LINUXKM_SIMD_IRQ */
return 0;
}
}
void restore_vector_registers_x86(void)
{
int processor_id = smp_processor_id();
if ((wolfcrypt_linuxkm_fpu_states == NULL)
#ifdef LINUXKM_SIMD_IRQ
|| (wolfcrypt_linuxkm_fpu_states[processor_id] == NULL)
#endif
)
{
pr_err("restore_vector_registers_x86 called for cpu id %d "
"with null context buffer.\n", processor_id);
return;
}
#ifdef LINUXKM_SIMD_IRQ
if (((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] == 0)
{
pr_err("restore_vector_registers_x86 called for cpu id %d "
"without saved context.\n", processor_id);
return;
}
if (--((unsigned char *)wolfcrypt_linuxkm_fpu_states[processor_id])[PAGE_SIZE-1] > 0) {
preempt_enable(); /* preempt_disable count will still be nonzero after this decrement. */
return;
}
if (am_in_hard_interrupt_handler()) {
#if LINUX_VERSION_CODE < KERNEL_VERSION(5, 14, 0)
copy_kernel_to_fpregs(wolfcrypt_linuxkm_fpu_states[processor_id]);
#elif LINUX_VERSION_CODE < KERNEL_VERSION(5, 16, 0)
__restore_fpregs_from_fpstate(wolfcrypt_linuxkm_fpu_states[processor_id],
xfeatures_mask_all);
#else
restore_fpregs_from_fpstate(wolfcrypt_linuxkm_fpu_states[processor_id],
fpu_kernel_cfg.max_features);
#endif
preempt_enable();
} else {
kernel_fpu_end();
}
#else /* !LINUXKM_SIMD_IRQ */
if (wolfcrypt_linuxkm_fpu_states[processor_id] == 0)
{
pr_err("restore_vector_registers_x86 called for cpu id %d "
"without saved context.\n", processor_id);
return;
}
if (--wolfcrypt_linuxkm_fpu_states[processor_id] > 0) {
preempt_enable(); /* preempt_disable count will still be nonzero after this decrement. */
return;
}
kernel_fpu_end();
#endif /* !LINUXKM_SIMD_IRQ */
return;
}
#endif /* WOLFSSL_LINUXKM_SIMD_X86 && WOLFSSL_LINUXKM_SIMD_X86_IRQ_ALLOWED */
#if defined(__PIE__) && (LINUX_VERSION_CODE >= KERNEL_VERSION(6, 1, 0))
/* needed in 6.1+ because show_free_areas() static definition in mm.h calls
* __show_free_areas(), which isn't exported (neither was show_free_areas()).
*/
void my__show_free_areas(
unsigned int flags,
nodemask_t *nodemask,
int max_zone_idx)
{
(void)flags;
(void)nodemask;
(void)max_zone_idx;
return;
}
#endif
|
