1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ================= 3 =================
4 KVM Lock Overview 4 KVM Lock Overview
5 ================= 5 =================
6 6
7 1. Acquisition Orders 7 1. Acquisition Orders
8 --------------------- 8 ---------------------
9 9
10 The acquisition orders for mutexes are as foll 10 The acquisition orders for mutexes are as follows:
11 11
12 - cpus_read_lock() is taken outside kvm_lock 12 - cpus_read_lock() is taken outside kvm_lock
13 13
14 - kvm_usage_lock is taken outside cpus_read_lo <<
15 <<
16 - kvm->lock is taken outside vcpu->mutex 14 - kvm->lock is taken outside vcpu->mutex
17 15
18 - kvm->lock is taken outside kvm->slots_lock a 16 - kvm->lock is taken outside kvm->slots_lock and kvm->irq_lock
19 17
20 - kvm->slots_lock is taken outside kvm->irq_lo 18 - kvm->slots_lock is taken outside kvm->irq_lock, though acquiring
21 them together is quite rare. 19 them together is quite rare.
22 20
23 - kvm->mn_active_invalidate_count ensures that 21 - kvm->mn_active_invalidate_count ensures that pairs of
24 invalidate_range_start() and invalidate_rang 22 invalidate_range_start() and invalidate_range_end() callbacks
25 use the same memslots array. kvm->slots_loc 23 use the same memslots array. kvm->slots_lock and kvm->slots_arch_lock
26 are taken on the waiting side when modifying 24 are taken on the waiting side when modifying memslots, so MMU notifiers
27 must not take either kvm->slots_lock or kvm- 25 must not take either kvm->slots_lock or kvm->slots_arch_lock.
28 26
29 cpus_read_lock() vs kvm_lock: <<
30 <<
31 - Taking cpus_read_lock() outside of kvm_lock <<
32 being the official ordering, as it is quite <<
33 cpus_read_lock() while holding kvm_lock. Us <<
34 e.g. avoid complex operations when possible. <<
35 <<
36 For SRCU: 27 For SRCU:
37 28
38 - ``synchronize_srcu(&kvm->srcu)`` is called i 29 - ``synchronize_srcu(&kvm->srcu)`` is called inside critical sections
39 for kvm->lock, vcpu->mutex and kvm->slots_lo 30 for kvm->lock, vcpu->mutex and kvm->slots_lock. These locks _cannot_
40 be taken inside a kvm->srcu read-side critic 31 be taken inside a kvm->srcu read-side critical section; that is, the
41 following is broken:: 32 following is broken::
42 33
43 srcu_read_lock(&kvm->srcu); 34 srcu_read_lock(&kvm->srcu);
44 mutex_lock(&kvm->slots_lock); 35 mutex_lock(&kvm->slots_lock);
45 36
46 - kvm->slots_arch_lock instead is released bef 37 - kvm->slots_arch_lock instead is released before the call to
47 ``synchronize_srcu()``. It _can_ therefore 38 ``synchronize_srcu()``. It _can_ therefore be taken inside a
48 kvm->srcu read-side critical section, for ex 39 kvm->srcu read-side critical section, for example while processing
49 a vmexit. 40 a vmexit.
50 41
51 On x86: 42 On x86:
52 43
53 - vcpu->mutex is taken outside kvm->arch.hyper 44 - vcpu->mutex is taken outside kvm->arch.hyperv.hv_lock and kvm->arch.xen.xen_lock
54 45
55 - kvm->arch.mmu_lock is an rwlock; critical se !! 46 - kvm->arch.mmu_lock is an rwlock. kvm->arch.tdp_mmu_pages_lock and
56 kvm->arch.tdp_mmu_pages_lock and kvm->arch.m !! 47 kvm->arch.mmu_unsync_pages_lock are taken inside kvm->arch.mmu_lock, and
57 also take kvm->arch.mmu_lock !! 48 cannot be taken without already holding kvm->arch.mmu_lock (typically with
>> 49 ``read_lock`` for the TDP MMU, thus the need for additional spinlocks).
58 50
59 Everything else is a leaf: no other lock is ta 51 Everything else is a leaf: no other lock is taken inside the critical
60 sections. 52 sections.
61 53
62 2. Exception 54 2. Exception
63 ------------ 55 ------------
64 56
65 Fast page fault: 57 Fast page fault:
66 58
67 Fast page fault is the fast path which fixes t 59 Fast page fault is the fast path which fixes the guest page fault out of
68 the mmu-lock on x86. Currently, the page fault 60 the mmu-lock on x86. Currently, the page fault can be fast in one of the
69 following two cases: 61 following two cases:
70 62
71 1. Access Tracking: The SPTE is not present, b 63 1. Access Tracking: The SPTE is not present, but it is marked for access
72 tracking. That means we need to restore the 64 tracking. That means we need to restore the saved R/X bits. This is
73 described in more detail later below. 65 described in more detail later below.
74 66
75 2. Write-Protection: The SPTE is present and t 67 2. Write-Protection: The SPTE is present and the fault is caused by
76 write-protect. That means we just need to c 68 write-protect. That means we just need to change the W bit of the spte.
77 69
78 What we use to avoid all the races is the Host 70 What we use to avoid all the races is the Host-writable bit and MMU-writable bit
79 on the spte: 71 on the spte:
80 72
81 - Host-writable means the gfn is writable in t 73 - Host-writable means the gfn is writable in the host kernel page tables and in
82 its KVM memslot. 74 its KVM memslot.
83 - MMU-writable means the gfn is writable in th 75 - MMU-writable means the gfn is writable in the guest's mmu and it is not
84 write-protected by shadow page write-protect 76 write-protected by shadow page write-protection.
85 77
86 On fast page fault path, we will use cmpxchg t 78 On fast page fault path, we will use cmpxchg to atomically set the spte W
87 bit if spte.HOST_WRITEABLE = 1 and spte.WRITE_ 79 bit if spte.HOST_WRITEABLE = 1 and spte.WRITE_PROTECT = 1, to restore the saved
88 R/X bits if for an access-traced spte, or both 80 R/X bits if for an access-traced spte, or both. This is safe because whenever
89 changing these bits can be detected by cmpxchg 81 changing these bits can be detected by cmpxchg.
90 82
91 But we need carefully check these cases: 83 But we need carefully check these cases:
92 84
93 1) The mapping from gfn to pfn 85 1) The mapping from gfn to pfn
94 86
95 The mapping from gfn to pfn may be changed sin 87 The mapping from gfn to pfn may be changed since we can only ensure the pfn
96 is not changed during cmpxchg. This is a ABA p 88 is not changed during cmpxchg. This is a ABA problem, for example, below case
97 will happen: 89 will happen:
98 90
99 +--------------------------------------------- 91 +------------------------------------------------------------------------+
100 | At the beginning:: 92 | At the beginning:: |
101 | 93 | |
102 | gpte = gfn1 94 | gpte = gfn1 |
103 | gfn1 is mapped to pfn1 on host 95 | gfn1 is mapped to pfn1 on host |
104 | spte is the shadow page table entry co 96 | spte is the shadow page table entry corresponding with gpte and |
105 | spte = pfn1 97 | spte = pfn1 |
106 +--------------------------------------------- 98 +------------------------------------------------------------------------+
107 | On fast page fault path: 99 | On fast page fault path: |
108 +------------------------------------+-------- 100 +------------------------------------+-----------------------------------+
109 | CPU 0: | CPU 1: 101 | CPU 0: | CPU 1: |
110 +------------------------------------+-------- 102 +------------------------------------+-----------------------------------+
111 | :: | 103 | :: | |
112 | | 104 | | |
113 | old_spte = *spte; | 105 | old_spte = *spte; | |
114 +------------------------------------+-------- 106 +------------------------------------+-----------------------------------+
115 | | pfn1 is 107 | | pfn1 is swapped out:: |
116 | | 108 | | |
117 | | spte 109 | | spte = 0; |
118 | | 110 | | |
119 | | pfn1 is 111 | | pfn1 is re-alloced for gfn2. |
120 | | 112 | | |
121 | | gpte is 113 | | gpte is changed to point to |
122 | | gfn2 by 114 | | gfn2 by the guest:: |
123 | | 115 | | |
124 | | spte 116 | | spte = pfn1; |
125 +------------------------------------+-------- 117 +------------------------------------+-----------------------------------+
126 | :: 118 | :: |
127 | 119 | |
128 | if (cmpxchg(spte, old_spte, old_spte+W) 120 | if (cmpxchg(spte, old_spte, old_spte+W) |
129 | mark_page_dirty(vcpu->kvm, gfn1) 121 | mark_page_dirty(vcpu->kvm, gfn1) |
130 | OOPS!!! 122 | OOPS!!! |
131 +--------------------------------------------- 123 +------------------------------------------------------------------------+
132 124
133 We dirty-log for gfn1, that means gfn2 is lost 125 We dirty-log for gfn1, that means gfn2 is lost in dirty-bitmap.
134 126
135 For direct sp, we can easily avoid it since th 127 For direct sp, we can easily avoid it since the spte of direct sp is fixed
136 to gfn. For indirect sp, we disabled fast pag 128 to gfn. For indirect sp, we disabled fast page fault for simplicity.
137 129
138 A solution for indirect sp could be to pin the 130 A solution for indirect sp could be to pin the gfn, for example via
139 gfn_to_pfn_memslot_atomic, before the cmpxchg. !! 131 kvm_vcpu_gfn_to_pfn_atomic, before the cmpxchg. After the pinning:
140 132
141 - We have held the refcount of pfn; that means 133 - We have held the refcount of pfn; that means the pfn can not be freed and
142 be reused for another gfn. 134 be reused for another gfn.
143 - The pfn is writable and therefore it cannot 135 - The pfn is writable and therefore it cannot be shared between different gfns
144 by KSM. 136 by KSM.
145 137
146 Then, we can ensure the dirty bitmaps is corre 138 Then, we can ensure the dirty bitmaps is correctly set for a gfn.
147 139
148 2) Dirty bit tracking 140 2) Dirty bit tracking
149 141
150 In the origin code, the spte can be fast updat 142 In the origin code, the spte can be fast updated (non-atomically) if the
151 spte is read-only and the Accessed bit has alr 143 spte is read-only and the Accessed bit has already been set since the
152 Accessed bit and Dirty bit can not be lost. 144 Accessed bit and Dirty bit can not be lost.
153 145
154 But it is not true after fast page fault since 146 But it is not true after fast page fault since the spte can be marked
155 writable between reading spte and updating spt 147 writable between reading spte and updating spte. Like below case:
156 148
157 +--------------------------------------------- 149 +------------------------------------------------------------------------+
158 | At the beginning:: 150 | At the beginning:: |
159 | 151 | |
160 | spte.W = 0 152 | spte.W = 0 |
161 | spte.Accessed = 1 153 | spte.Accessed = 1 |
162 +------------------------------------+-------- 154 +------------------------------------+-----------------------------------+
163 | CPU 0: | CPU 1: 155 | CPU 0: | CPU 1: |
164 +------------------------------------+-------- 156 +------------------------------------+-----------------------------------+
165 | In mmu_spte_clear_track_bits():: | 157 | In mmu_spte_clear_track_bits():: | |
166 | | 158 | | |
167 | old_spte = *spte; | 159 | old_spte = *spte; | |
168 | | 160 | | |
169 | | 161 | | |
170 | /* 'if' condition is satisfied. */| 162 | /* 'if' condition is satisfied. */| |
171 | if (old_spte.Accessed == 1 && | 163 | if (old_spte.Accessed == 1 && | |
172 | old_spte.W == 0) | 164 | old_spte.W == 0) | |
173 | spte = 0ull; | 165 | spte = 0ull; | |
174 +------------------------------------+-------- 166 +------------------------------------+-----------------------------------+
175 | | on fast 167 | | on fast page fault path:: |
176 | | 168 | | |
177 | | spte 169 | | spte.W = 1 |
178 | | 170 | | |
179 | | memory 171 | | memory write on the spte:: |
180 | | 172 | | |
181 | | spte 173 | | spte.Dirty = 1 |
182 +------------------------------------+-------- 174 +------------------------------------+-----------------------------------+
183 | :: | 175 | :: | |
184 | | 176 | | |
185 | else | 177 | else | |
186 | old_spte = xchg(spte, 0ull) | 178 | old_spte = xchg(spte, 0ull) | |
187 | if (old_spte.Accessed == 1) | 179 | if (old_spte.Accessed == 1) | |
188 | kvm_set_pfn_accessed(spte.pfn);| 180 | kvm_set_pfn_accessed(spte.pfn);| |
189 | if (old_spte.Dirty == 1) | 181 | if (old_spte.Dirty == 1) | |
190 | kvm_set_pfn_dirty(spte.pfn); | 182 | kvm_set_pfn_dirty(spte.pfn); | |
191 | OOPS!!! | 183 | OOPS!!! | |
192 +------------------------------------+-------- 184 +------------------------------------+-----------------------------------+
193 185
194 The Dirty bit is lost in this case. 186 The Dirty bit is lost in this case.
195 187
196 In order to avoid this kind of issue, we alway 188 In order to avoid this kind of issue, we always treat the spte as "volatile"
197 if it can be updated out of mmu-lock [see spte 189 if it can be updated out of mmu-lock [see spte_has_volatile_bits()]; it means
198 the spte is always atomically updated in this 190 the spte is always atomically updated in this case.
199 191
200 3) flush tlbs due to spte updated 192 3) flush tlbs due to spte updated
201 193
202 If the spte is updated from writable to read-o 194 If the spte is updated from writable to read-only, we should flush all TLBs,
203 otherwise rmap_write_protect will find a read- 195 otherwise rmap_write_protect will find a read-only spte, even though the
204 writable spte might be cached on a CPU's TLB. 196 writable spte might be cached on a CPU's TLB.
205 197
206 As mentioned before, the spte can be updated t 198 As mentioned before, the spte can be updated to writable out of mmu-lock on
207 fast page fault path. In order to easily audit 199 fast page fault path. In order to easily audit the path, we see if TLBs needing
208 to be flushed caused this reason in mmu_spte_u 200 to be flushed caused this reason in mmu_spte_update() since this is a common
209 function to update spte (present -> present). 201 function to update spte (present -> present).
210 202
211 Since the spte is "volatile" if it can be upda 203 Since the spte is "volatile" if it can be updated out of mmu-lock, we always
212 atomically update the spte and the race caused 204 atomically update the spte and the race caused by fast page fault can be avoided.
213 See the comments in spte_has_volatile_bits() a 205 See the comments in spte_has_volatile_bits() and mmu_spte_update().
214 206
215 Lockless Access Tracking: 207 Lockless Access Tracking:
216 208
217 This is used for Intel CPUs that are using EPT 209 This is used for Intel CPUs that are using EPT but do not support the EPT A/D
218 bits. In this case, PTEs are tagged as A/D dis 210 bits. In this case, PTEs are tagged as A/D disabled (using ignored bits), and
219 when the KVM MMU notifier is called to track a 211 when the KVM MMU notifier is called to track accesses to a page (via
220 kvm_mmu_notifier_clear_flush_young), it marks 212 kvm_mmu_notifier_clear_flush_young), it marks the PTE not-present in hardware
221 by clearing the RWX bits in the PTE and storin 213 by clearing the RWX bits in the PTE and storing the original R & X bits in more
222 unused/ignored bits. When the VM tries to acce 214 unused/ignored bits. When the VM tries to access the page later on, a fault is
223 generated and the fast page fault mechanism de 215 generated and the fast page fault mechanism described above is used to
224 atomically restore the PTE to a Present state. 216 atomically restore the PTE to a Present state. The W bit is not saved when the
225 PTE is marked for access tracking and during r 217 PTE is marked for access tracking and during restoration to the Present state,
226 the W bit is set depending on whether or not i 218 the W bit is set depending on whether or not it was a write access. If it
227 wasn't, then the W bit will remain clear until 219 wasn't, then the W bit will remain clear until a write access happens, at which
228 time it will be set using the Dirty tracking m 220 time it will be set using the Dirty tracking mechanism described above.
229 221
230 3. Reference 222 3. Reference
231 ------------ 223 ------------
232 224
233 ``kvm_lock`` 225 ``kvm_lock``
234 ^^^^^^^^^^^^ 226 ^^^^^^^^^^^^
235 227
236 :Type: mutex 228 :Type: mutex
237 :Arch: any 229 :Arch: any
238 :Protects: - vm_list 230 :Protects: - vm_list
239 !! 231 - kvm_usage_count
240 ``kvm_usage_lock`` <<
241 ^^^^^^^^^^^^^^^^^^ <<
242 <<
243 :Type: mutex <<
244 :Arch: any <<
245 :Protects: - kvm_usage_count <<
246 - hardware virtualization enab 232 - hardware virtualization enable/disable
247 :Comment: Exists to allow taking cpus_re !! 233 :Comment: KVM also disables CPU hotplug via cpus_read_lock() during
248 protected, which simplifies th !! 234 enable/disable.
249 235
250 ``kvm->mn_invalidate_lock`` 236 ``kvm->mn_invalidate_lock``
251 ^^^^^^^^^^^^^^^^^^^^^^^^^^^ 237 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
252 238
253 :Type: spinlock_t 239 :Type: spinlock_t
254 :Arch: any 240 :Arch: any
255 :Protects: mn_active_invalidate_count, mn 241 :Protects: mn_active_invalidate_count, mn_memslots_update_rcuwait
256 242
257 ``kvm_arch::tsc_write_lock`` 243 ``kvm_arch::tsc_write_lock``
258 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 244 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
259 245
260 :Type: raw_spinlock_t 246 :Type: raw_spinlock_t
261 :Arch: x86 247 :Arch: x86
262 :Protects: - kvm_arch::{last_tsc_write,la 248 :Protects: - kvm_arch::{last_tsc_write,last_tsc_nsec,last_tsc_offset}
263 - tsc offset in vmcb 249 - tsc offset in vmcb
264 :Comment: 'raw' because updating the tsc 250 :Comment: 'raw' because updating the tsc offsets must not be preempted.
265 251
266 ``kvm->mmu_lock`` 252 ``kvm->mmu_lock``
267 ^^^^^^^^^^^^^^^^^ 253 ^^^^^^^^^^^^^^^^^
268 :Type: spinlock_t or rwlock_t 254 :Type: spinlock_t or rwlock_t
269 :Arch: any 255 :Arch: any
270 :Protects: -shadow page/shadow tlb entry 256 :Protects: -shadow page/shadow tlb entry
271 :Comment: it is a spinlock since it is u 257 :Comment: it is a spinlock since it is used in mmu notifier.
272 258
273 ``kvm->srcu`` 259 ``kvm->srcu``
274 ^^^^^^^^^^^^^ 260 ^^^^^^^^^^^^^
275 :Type: srcu lock 261 :Type: srcu lock
276 :Arch: any 262 :Arch: any
277 :Protects: - kvm->memslots 263 :Protects: - kvm->memslots
278 - kvm->buses 264 - kvm->buses
279 :Comment: The srcu read lock must be hel 265 :Comment: The srcu read lock must be held while accessing memslots (e.g.
280 when using gfn_to_* functions) 266 when using gfn_to_* functions) and while accessing in-kernel
281 MMIO/PIO address->device struc 267 MMIO/PIO address->device structure mapping (kvm->buses).
282 The srcu index can be stored i 268 The srcu index can be stored in kvm_vcpu->srcu_idx per vcpu
283 if it is needed by multiple fu 269 if it is needed by multiple functions.
284 270
285 ``kvm->slots_arch_lock`` 271 ``kvm->slots_arch_lock``
286 ^^^^^^^^^^^^^^^^^^^^^^^^ 272 ^^^^^^^^^^^^^^^^^^^^^^^^
287 :Type: mutex 273 :Type: mutex
288 :Arch: any (only needed on x86 though 274 :Arch: any (only needed on x86 though)
289 :Protects: any arch-specific fields of me 275 :Protects: any arch-specific fields of memslots that have to be modified
290 in a ``kvm->srcu`` read-side c 276 in a ``kvm->srcu`` read-side critical section.
291 :Comment: must be held before reading th 277 :Comment: must be held before reading the pointer to the current memslots,
292 until after all changes to the 278 until after all changes to the memslots are complete
293 279
294 ``wakeup_vcpus_on_cpu_lock`` 280 ``wakeup_vcpus_on_cpu_lock``
295 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 281 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
296 :Type: spinlock_t 282 :Type: spinlock_t
297 :Arch: x86 283 :Arch: x86
298 :Protects: wakeup_vcpus_on_cpu 284 :Protects: wakeup_vcpus_on_cpu
299 :Comment: This is a per-CPU lock and it 285 :Comment: This is a per-CPU lock and it is used for VT-d posted-interrupts.
300 When VT-d posted-interrupts ar 286 When VT-d posted-interrupts are supported and the VM has assigned
301 devices, we put the blocked vC 287 devices, we put the blocked vCPU on the list blocked_vcpu_on_cpu
302 protected by blocked_vcpu_on_c 288 protected by blocked_vcpu_on_cpu_lock. When VT-d hardware issues
303 wakeup notification event sinc 289 wakeup notification event since external interrupts from the
304 assigned devices happens, we w 290 assigned devices happens, we will find the vCPU on the list to
305 wakeup. 291 wakeup.
306 292
307 ``vendor_module_lock`` 293 ``vendor_module_lock``
308 ^^^^^^^^^^^^^^^^^^^^^^ !! 294 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
309 :Type: mutex 295 :Type: mutex
310 :Arch: x86 296 :Arch: x86
311 :Protects: loading a vendor module (kvm_a 297 :Protects: loading a vendor module (kvm_amd or kvm_intel)
312 :Comment: Exists because using kvm_lock !! 298 :Comment: Exists because using kvm_lock leads to deadlock. cpu_hotplug_lock is
313 in notifiers, e.g. __kvmclock_cpufreq_noti !! 299 taken outside of kvm_lock, e.g. in KVM's CPU online/offline callbacks, and
314 cpu_hotplug_lock is held, e.g. from cpufre !! 300 many operations need to take cpu_hotplug_lock when loading a vendor module,
315 operations need to take cpu_hotplug_lock w !! 301 e.g. updating static calls.
316 updating static calls. <<
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.