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
|
/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License, version 2, as
* published by the Free Software Foundation.
*
* This program 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, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* Copyright SUSE Linux Products GmbH 2009
*
* Authors: Alexander Graf <agraf@suse.de>
*/
#include <asm/ppc_asm.h>
#include <asm/kvm_asm.h>
#include <asm/reg.h>
#include <asm/page.h>
#include <asm/asm-offsets.h>
#include <asm/exception-64s.h>
#define KVMPPC_HANDLE_EXIT .kvmppc_handle_exit
#define ULONG_SIZE 8
#define VCPU_GPR(n) (VCPU_GPRS + (n * ULONG_SIZE))
.macro DISABLE_INTERRUPTS
mfmsr r0
rldicl r0,r0,48,1
rotldi r0,r0,16
mtmsrd r0,1
.endm
#define VCPU_LOAD_NVGPRS(vcpu) \
ld r14, VCPU_GPR(r14)(vcpu); \
ld r15, VCPU_GPR(r15)(vcpu); \
ld r16, VCPU_GPR(r16)(vcpu); \
ld r17, VCPU_GPR(r17)(vcpu); \
ld r18, VCPU_GPR(r18)(vcpu); \
ld r19, VCPU_GPR(r19)(vcpu); \
ld r20, VCPU_GPR(r20)(vcpu); \
ld r21, VCPU_GPR(r21)(vcpu); \
ld r22, VCPU_GPR(r22)(vcpu); \
ld r23, VCPU_GPR(r23)(vcpu); \
ld r24, VCPU_GPR(r24)(vcpu); \
ld r25, VCPU_GPR(r25)(vcpu); \
ld r26, VCPU_GPR(r26)(vcpu); \
ld r27, VCPU_GPR(r27)(vcpu); \
ld r28, VCPU_GPR(r28)(vcpu); \
ld r29, VCPU_GPR(r29)(vcpu); \
ld r30, VCPU_GPR(r30)(vcpu); \
ld r31, VCPU_GPR(r31)(vcpu); \
/*****************************************************************************
* *
* Guest entry / exit code that is in kernel module memory (highmem) *
* *
****************************************************************************/
/* Registers:
* r3: kvm_run pointer
* r4: vcpu pointer
*/
_GLOBAL(__kvmppc_vcpu_entry)
kvm_start_entry:
/* Write correct stack frame */
mflr r0
std r0,16(r1)
/* Save host state to the stack */
stdu r1, -SWITCH_FRAME_SIZE(r1)
/* Save r3 (kvm_run) and r4 (vcpu) */
SAVE_2GPRS(3, r1)
/* Save non-volatile registers (r14 - r31) */
SAVE_NVGPRS(r1)
/* Save LR */
std r0, _LINK(r1)
/* Load non-volatile guest state from the vcpu */
VCPU_LOAD_NVGPRS(r4)
/* Save R1/R2 in the PACA */
std r1, PACA_KVM_HOST_R1(r13)
std r2, PACA_KVM_HOST_R2(r13)
/* XXX swap in/out on load? */
ld r3, VCPU_HIGHMEM_HANDLER(r4)
std r3, PACA_KVM_VMHANDLER(r13)
kvm_start_lightweight:
ld r9, VCPU_PC(r4) /* r9 = vcpu->arch.pc */
ld r10, VCPU_SHADOW_MSR(r4) /* r10 = vcpu->arch.shadow_msr */
/* Load some guest state in the respective registers */
ld r5, VCPU_CTR(r4) /* r5 = vcpu->arch.ctr */
/* will be swapped in by rmcall */
ld r3, VCPU_LR(r4) /* r3 = vcpu->arch.lr */
mtlr r3 /* LR = r3 */
DISABLE_INTERRUPTS
/* Some guests may need to have dcbz set to 32 byte length.
*
* Usually we ensure that by patching the guest's instructions
* to trap on dcbz and emulate it in the hypervisor.
*
* If we can, we should tell the CPU to use 32 byte dcbz though,
* because that's a lot faster.
*/
ld r3, VCPU_HFLAGS(r4)
rldicl. r3, r3, 0, 63 /* CR = ((r3 & 1) == 0) */
beq no_dcbz32_on
mfspr r3,SPRN_HID5
ori r3, r3, 0x80 /* XXX HID5_dcbz32 = 0x80 */
mtspr SPRN_HID5,r3
no_dcbz32_on:
ld r6, VCPU_RMCALL(r4)
mtctr r6
ld r3, VCPU_TRAMPOLINE_ENTER(r4)
LOAD_REG_IMMEDIATE(r4, MSR_KERNEL & ~(MSR_IR | MSR_DR))
/* Jump to SLB patching handlder and into our guest */
bctr
/*
* This is the handler in module memory. It gets jumped at from the
* lowmem trampoline code, so it's basically the guest exit code.
*
*/
.global kvmppc_handler_highmem
kvmppc_handler_highmem:
/*
* Register usage at this point:
*
* R0 = guest last inst
* R1 = host R1
* R2 = host R2
* R3 = guest PC
* R4 = guest MSR
* R5 = guest DAR
* R6 = guest DSISR
* R13 = PACA
* PACA.KVM.* = guest *
*
*/
/* R7 = vcpu */
ld r7, GPR4(r1)
/* Now save the guest state */
stw r0, VCPU_LAST_INST(r7)
std r3, VCPU_PC(r7)
std r4, VCPU_SHADOW_MSR(r7)
std r5, VCPU_FAULT_DEAR(r7)
std r6, VCPU_FAULT_DSISR(r7)
ld r5, VCPU_HFLAGS(r7)
rldicl. r5, r5, 0, 63 /* CR = ((r5 & 1) == 0) */
beq no_dcbz32_off
mfspr r5,SPRN_HID5
rldimi r5,r5,6,56
mtspr SPRN_HID5,r5
no_dcbz32_off:
std r14, VCPU_GPR(r14)(r7)
std r15, VCPU_GPR(r15)(r7)
std r16, VCPU_GPR(r16)(r7)
std r17, VCPU_GPR(r17)(r7)
std r18, VCPU_GPR(r18)(r7)
std r19, VCPU_GPR(r19)(r7)
std r20, VCPU_GPR(r20)(r7)
std r21, VCPU_GPR(r21)(r7)
std r22, VCPU_GPR(r22)(r7)
std r23, VCPU_GPR(r23)(r7)
std r24, VCPU_GPR(r24)(r7)
std r25, VCPU_GPR(r25)(r7)
std r26, VCPU_GPR(r26)(r7)
std r27, VCPU_GPR(r27)(r7)
std r28, VCPU_GPR(r28)(r7)
std r29, VCPU_GPR(r29)(r7)
std r30, VCPU_GPR(r30)(r7)
std r31, VCPU_GPR(r31)(r7)
/* Save guest CTR */
mfctr r5
std r5, VCPU_CTR(r7)
/* Save guest LR */
mflr r5
std r5, VCPU_LR(r7)
/* Restore host msr -> SRR1 */
ld r6, VCPU_HOST_MSR(r7)
/*
* For some interrupts, we need to call the real Linux
* handler, so it can do work for us. This has to happen
* as if the interrupt arrived from the kernel though,
* so let's fake it here where most state is restored.
*
* Call Linux for hardware interrupts/decrementer
* r3 = address of interrupt handler (exit reason)
*/
cmpwi r12, BOOK3S_INTERRUPT_EXTERNAL
beq call_linux_handler
cmpwi r12, BOOK3S_INTERRUPT_DECREMENTER
beq call_linux_handler
/* Back to EE=1 */
mtmsr r6
b kvm_return_point
call_linux_handler:
/*
* If we land here we need to jump back to the handler we
* came from.
*
* We have a page that we can access from real mode, so let's
* jump back to that and use it as a trampoline to get back into the
* interrupt handler!
*
* R3 still contains the exit code,
* R5 VCPU_HOST_RETIP and
* R6 VCPU_HOST_MSR
*/
/* Restore host IP -> SRR0 */
ld r5, VCPU_HOST_RETIP(r7)
/* XXX Better move to a safe function?
* What if we get an HTAB flush in between mtsrr0 and mtsrr1? */
mtlr r12
ld r4, VCPU_TRAMPOLINE_LOWMEM(r7)
mtsrr0 r4
LOAD_REG_IMMEDIATE(r3, MSR_KERNEL & ~(MSR_IR | MSR_DR))
mtsrr1 r3
RFI
.global kvm_return_point
kvm_return_point:
/* Jump back to lightweight entry if we're supposed to */
/* go back into the guest */
/* Pass the exit number as 3rd argument to kvmppc_handle_exit */
mr r5, r12
/* Restore r3 (kvm_run) and r4 (vcpu) */
REST_2GPRS(3, r1)
bl KVMPPC_HANDLE_EXIT
/* If RESUME_GUEST, get back in the loop */
cmpwi r3, RESUME_GUEST
beq kvm_loop_lightweight
cmpwi r3, RESUME_GUEST_NV
beq kvm_loop_heavyweight
kvm_exit_loop:
ld r4, _LINK(r1)
mtlr r4
/* Restore non-volatile host registers (r14 - r31) */
REST_NVGPRS(r1)
addi r1, r1, SWITCH_FRAME_SIZE
blr
kvm_loop_heavyweight:
ld r4, _LINK(r1)
std r4, (16 + SWITCH_FRAME_SIZE)(r1)
/* Load vcpu and cpu_run */
REST_2GPRS(3, r1)
/* Load non-volatile guest state from the vcpu */
VCPU_LOAD_NVGPRS(r4)
/* Jump back into the beginning of this function */
b kvm_start_lightweight
kvm_loop_lightweight:
/* We'll need the vcpu pointer */
REST_GPR(4, r1)
/* Jump back into the beginning of this function */
b kvm_start_lightweight
|
