€•ŸÛŒsphinx.addnodes”Œdocument”“”)”}”(Œ rawsource”Œ”Œchildren”]”(Œtranslations”Œ LanguagesNode”“”)”}”(hhh]”(hŒpending_xref”“”)”}”(hhh]”Œdocutils.nodes”ŒText”“”ŒChinese (Simplified)”…””}”Œparent”hsbaŒ attributes”}”(Œids”]”Œclasses”]”Œnames”]”Œdupnames”]”Œbackrefs”]”Œ refdomain”Œstd”Œreftype”Œdoc”Œ reftarget”Œ$/translations/zh_CN/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuŒtagname”hhhubh)”}”(hhh]”hŒChinese (Traditional)”…””}”hh2sbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/zh_TW/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒItalian”…””}”hhFsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/it_IT/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒJapanese”…””}”hhZsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/ja_JP/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒKorean”…””}”hhnsbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/ko_KR/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒPortuguese (Brazilian)”…””}”hh‚sbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/pt_BR/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubh)”}”(hhh]”hŒSpanish”…””}”hh–sbah}”(h]”h ]”h"]”h$]”h&]”Œ refdomain”h)Œreftype”h+Œ reftarget”Œ$/translations/sp_SP/virt/kvm/locking”Œmodname”NŒ classname”NŒrefexplicit”ˆuh1hhhubeh}”(h]”h ]”h"]”h$]”h&]”Œcurrent_language”ŒEnglish”uh1h hhŒ _document”hŒsource”NŒline”NubhŒcomment”“”)”}”(hŒ SPDX-License-Identifier: GPL-2.0”h]”hŒ SPDX-License-Identifier: GPL-2.0”…””}”hh·sbah}”(h]”h ]”h"]”h$]”h&]”Œ xml:space”Œpreserve”uh1hµhhh²hh³Œ>/var/lib/git/docbuild/linux/Documentation/virt/kvm/locking.rst”h´KubhŒsection”“”)”}”(hhh]”(hŒtitle”“”)”}”(hŒKVM Lock Overview”h]”hŒKVM Lock Overview”…””}”(hhÏh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhhÊh²hh³hÇh´KubhÉ)”}”(hhh]”(hÎ)”}”(hŒ1. Acquisition Orders”h]”hŒ1. Acquisition Orders”…””}”(hhàh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhhÝh²hh³hÇh´KubhŒ paragraph”“”)”}”(hŒ2The acquisition orders for mutexes are as follows:”h]”hŒ2The acquisition orders for mutexes are as follows:”…””}”(hhðh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K hhÝh²hubhŒbullet_list”“”)”}”(hhh]”(hŒ list_item”“”)”}”(hŒ+cpus_read_lock() is taken outside kvm_lock ”h]”hï)”}”(hŒ*cpus_read_lock() is taken outside kvm_lock”h]”hŒ*cpus_read_lock() is taken outside kvm_lock”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hŒ1kvm_usage_lock is taken outside cpus_read_lock() ”h]”hï)”}”(hŒ0kvm_usage_lock is taken outside cpus_read_lock()”h]”hŒ0kvm_usage_lock is taken outside cpus_read_lock()”…””}”(hj!h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hŒ'kvm->lock is taken outside vcpu->mutex ”h]”hï)”}”(hŒ&kvm->lock is taken outside vcpu->mutex”h]”hŒ&kvm->lock is taken outside vcpu->mutex”…””}”(hj9h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khj5ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hŒ=kvm->lock is taken outside kvm->slots_lock and kvm->irq_lock ”h]”hï)”}”(hŒlock is taken outside kvm->slots_lock and kvm->irq_lock”h]”hŒlock is taken outside kvm->slots_lock and kvm->irq_lock”…””}”(hjQh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KhjMubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hŒFvcpu->mutex is taken outside kvm->slots_lock and kvm->slots_arch_lock ”h]”hï)”}”(hŒEvcpu->mutex is taken outside kvm->slots_lock and kvm->slots_arch_lock”h]”hŒEvcpu->mutex is taken outside kvm->slots_lock and kvm->slots_arch_lock”…””}”(hjih²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khjeubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hŒ^kvm->slots_lock is taken outside kvm->irq_lock, though acquiring them together is quite rare. ”h]”hï)”}”(hŒ]kvm->slots_lock is taken outside kvm->irq_lock, though acquiring them together is quite rare.”h]”hŒ]kvm->slots_lock is taken outside kvm->irq_lock, though acquiring them together is quite rare.”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khj}ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubj)”}”(hXAkvm->mn_active_invalidate_count ensures that pairs of invalidate_range_start() and invalidate_range_end() callbacks use the same memslots array. kvm->slots_lock and kvm->slots_arch_lock are taken on the waiting side when modifying memslots, so MMU notifiers must not take either kvm->slots_lock or kvm->slots_arch_lock. ”h]”hï)”}”(hX@kvm->mn_active_invalidate_count ensures that pairs of invalidate_range_start() and invalidate_range_end() callbacks use the same memslots array. kvm->slots_lock and kvm->slots_arch_lock are taken on the waiting side when modifying memslots, so MMU notifiers must not take either kvm->slots_lock or kvm->slots_arch_lock.”h]”hX@kvm->mn_active_invalidate_count ensures that pairs of invalidate_range_start() and invalidate_range_end() callbacks use the same memslots array. kvm->slots_lock and kvm->slots_arch_lock are taken on the waiting side when modifying memslots, so MMU notifiers must not take either kvm->slots_lock or kvm->slots_arch_lock.”…””}”(hj™h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Khj•ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjh²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”Œbullet”Œ-”uh1hþh³hÇh´KhhÝh²hubhï)”}”(hŒcpus_read_lock() vs kvm_lock:”h]”hŒcpus_read_lock() vs kvm_lock:”…””}”(hjµh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KhhÝh²hubhÿ)”}”(hhh]”j)”}”(hX Taking cpus_read_lock() outside of kvm_lock is problematic, despite that being the official ordering, as it is quite easy to unknowingly trigger cpus_read_lock() while holding kvm_lock. Use caution when walking vm_list, e.g. avoid complex operations when possible. ”h]”hï)”}”(hX Taking cpus_read_lock() outside of kvm_lock is problematic, despite that being the official ordering, as it is quite easy to unknowingly trigger cpus_read_lock() while holding kvm_lock. Use caution when walking vm_list, e.g. avoid complex operations when possible.”h]”hX Taking cpus_read_lock() outside of kvm_lock is problematic, despite that being the official ordering, as it is quite easy to unknowingly trigger cpus_read_lock() while holding kvm_lock. Use caution when walking vm_list, e.g. avoid complex operations when possible.”…””}”(hjÊh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K!hjÆubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÃh²hh³hÇh´Nubah}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´K!hhÝh²hubhï)”}”(hŒ For SRCU:”h]”hŒ For SRCU:”…””}”(hjäh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K&hhÝh²hubhÿ)”}”(hhh]”(j)”}”(hX(``synchronize_srcu(&kvm->srcu)`` is called inside critical sections for kvm->lock, vcpu->mutex and kvm->slots_lock. These locks _cannot_ be taken inside a kvm->srcu read-side critical section; that is, the following is broken:: srcu_read_lock(&kvm->srcu); mutex_lock(&kvm->slots_lock); ”h]”(hï)”}”(hŒä``synchronize_srcu(&kvm->srcu)`` is called inside critical sections for kvm->lock, vcpu->mutex and kvm->slots_lock. These locks _cannot_ be taken inside a kvm->srcu read-side critical section; that is, the following is broken::”h]”(hŒliteral”“”)”}”(hŒ ``synchronize_srcu(&kvm->srcu)``”h]”hŒsynchronize_srcu(&kvm->srcu)”…””}”(hjÿh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjùubhŒÃ is called inside critical sections for kvm->lock, vcpu->mutex and kvm->slots_lock. These locks _cannot_ be taken inside a kvm->srcu read-side critical section; that is, the following is broken:”…””}”(hjùh²hh³Nh´Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K(hjõubhŒ literal_block”“”)”}”(hŒ9srcu_read_lock(&kvm->srcu); mutex_lock(&kvm->slots_lock);”h]”hŒ9srcu_read_lock(&kvm->srcu); mutex_lock(&kvm->slots_lock);”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K-hjõubeh}”(h]”h ]”h"]”h$]”h&]”uh1jhjòh²hh³hÇh´Nubj)”}”(hŒÇkvm->slots_arch_lock instead is released before the call to ``synchronize_srcu()``. It _can_ therefore be taken inside a kvm->srcu read-side critical section, for example while processing a vmexit. ”h]”hï)”}”(hŒÆkvm->slots_arch_lock instead is released before the call to ``synchronize_srcu()``. It _can_ therefore be taken inside a kvm->srcu read-side critical section, for example while processing a vmexit.”h]”(hŒslots_arch_lock instead is released before the call to ”…””}”(hj1h²hh³Nh´Nubjþ)”}”(hŒ``synchronize_srcu()``”h]”hŒsynchronize_srcu()”…””}”(hj9h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhj1ubhŒt. It _can_ therefore be taken inside a kvm->srcu read-side critical section, for example while processing a vmexit.”…””}”(hj1h²hh³Nh´Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K0hj-ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjòh²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´K(hhÝh²hubhï)”}”(hŒOn x86:”h]”hŒOn x86:”…””}”(hj]h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K5hhÝh²hubhÿ)”}”(hhh]”(j)”}”(hŒQvcpu->mutex is taken outside kvm->arch.hyperv.hv_lock and kvm->arch.xen.xen_lock ”h]”hï)”}”(hŒPvcpu->mutex is taken outside kvm->arch.hyperv.hv_lock and kvm->arch.xen.xen_lock”h]”hŒPvcpu->mutex is taken outside kvm->arch.hyperv.hv_lock and kvm->arch.xen.xen_lock”…””}”(hjrh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K7hjnubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjkh²hh³hÇh´Nubj)”}”(hŒškvm->arch.mmu_lock is an rwlock; critical sections for kvm->arch.tdp_mmu_pages_lock and kvm->arch.mmu_unsync_pages_lock must also take kvm->arch.mmu_lock ”h]”hï)”}”(hŒ™kvm->arch.mmu_lock is an rwlock; critical sections for kvm->arch.tdp_mmu_pages_lock and kvm->arch.mmu_unsync_pages_lock must also take kvm->arch.mmu_lock”h]”hŒ™kvm->arch.mmu_lock is an rwlock; critical sections for kvm->arch.tdp_mmu_pages_lock and kvm->arch.mmu_unsync_pages_lock must also take kvm->arch.mmu_lock”…””}”(hjŠh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K9hj†ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjkh²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´K7hhÝh²hubhï)”}”(hŒOEverything else is a leaf: no other lock is taken inside the critical sections.”h]”hŒOEverything else is a leaf: no other lock is taken inside the critical sections.”…””}”(hj¤h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K=hhÝh²hubeh}”(h]”Œacquisition-orders”ah ]”h"]”Œ1. acquisition orders”ah$]”h&]”uh1hÈhhÊh²hh³hÇh´KubhÉ)”}”(hhh]”(hÎ)”}”(hŒ2. Exception”h]”hŒ2. Exception”…””}”(hj½h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjºh²hh³hÇh´KAubhï)”}”(hŒFast page fault:”h]”hŒFast page fault:”…””}”(hjËh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KChjºh²hubhï)”}”(hŒ¦Fast page fault is the fast path which fixes the guest page fault out of the mmu-lock on x86. Currently, the page fault can be fast in one of the following two cases:”h]”hŒ¦Fast page fault is the fast path which fixes the guest page fault out of the mmu-lock on x86. Currently, the page fault can be fast in one of the following two cases:”…””}”(hjÙh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KEhjºh²hubhŒenumerated_list”“”)”}”(hhh]”(j)”}”(hŒ°Access Tracking: The SPTE is not present, but it is marked for access tracking. That means we need to restore the saved R/X bits. This is described in more detail later below. ”h]”hï)”}”(hŒ¯Access Tracking: The SPTE is not present, but it is marked for access tracking. That means we need to restore the saved R/X bits. This is described in more detail later below.”h]”hŒ¯Access Tracking: The SPTE is not present, but it is marked for access tracking. That means we need to restore the saved R/X bits. This is described in more detail later below.”…””}”(hjðh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KIhjìubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjéh²hh³hÇh´Nubj)”}”(hŒ‰Write-Protection: The SPTE is present and the fault is caused by write-protect. That means we just need to change the W bit of the spte. ”h]”hï)”}”(hŒˆWrite-Protection: The SPTE is present and the fault is caused by write-protect. That means we just need to change the W bit of the spte.”h]”hŒˆWrite-Protection: The SPTE is present and the fault is caused by write-protect. That means we just need to change the W bit of the spte.”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KMhjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjéh²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”Œenumtype”Œarabic”Œprefix”hŒsuffix”Œ.”uh1jçhjºh²hh³hÇh´KIubhï)”}”(hŒ]What we use to avoid all the races is the Host-writable bit and MMU-writable bit on the spte:”h]”hŒ]What we use to avoid all the races is the Host-writable bit and MMU-writable bit on the spte:”…””}”(hj'h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KPhjºh²hubhÿ)”}”(hhh]”(j)”}”(hŒ^Host-writable means the gfn is writable in the host kernel page tables and in its KVM memslot.”h]”hï)”}”(hŒ^Host-writable means the gfn is writable in the host kernel page tables and in its KVM memslot.”h]”hŒ^Host-writable means the gfn is writable in the host kernel page tables and in its KVM memslot.”…””}”(hj<h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KShj8ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj5h²hh³hÇh´Nubj)”}”(hŒyMMU-writable means the gfn is writable in the guest's mmu and it is not write-protected by shadow page write-protection. ”h]”hï)”}”(hŒxMMU-writable means the gfn is writable in the guest's mmu and it is not write-protected by shadow page write-protection.”h]”hŒzMMU-writable means the gfn is writable in the guestâ€™s mmu and it is not write-protected by shadow page write-protection.”…””}”(hjTh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KUhjPubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj5h²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´KShjºh²hubhï)”}”(hXOn fast page fault path, we will use cmpxchg to atomically set the spte W bit if spte.HOST_WRITEABLE = 1 and spte.WRITE_PROTECT = 1, to restore the saved R/X bits if for an access-traced spte, or both. This is safe because whenever changing these bits can be detected by cmpxchg.”h]”hXOn fast page fault path, we will use cmpxchg to atomically set the spte W bit if spte.HOST_WRITEABLE = 1 and spte.WRITE_PROTECT = 1, to restore the saved R/X bits if for an access-traced spte, or both. This is safe because whenever changing these bits can be detected by cmpxchg.”…””}”(hjnh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KXhjºh²hubhï)”}”(hŒ(But we need carefully check these cases:”h]”hŒ(But we need carefully check these cases:”…””}”(hj|h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K]hjºh²hubjè)”}”(hhh]”j)”}”(hŒThe mapping from gfn to pfn ”h]”hï)”}”(hŒThe mapping from gfn to pfn”h]”hŒThe mapping from gfn to pfn”…””}”(hj‘h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K_hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjŠh²hh³hÇh´Nubah}”(h]”h ]”h"]”h$]”h&]”j"j#j$hj%Œ)”uh1jçhjºh²hh³hÇh´K_ubhï)”}”(hŒ¦The mapping from gfn to pfn may be changed since we can only ensure the pfn is not changed during cmpxchg. This is a ABA problem, for example, below case will happen:”h]”hŒ¦The mapping from gfn to pfn may be changed since we can only ensure the pfn is not changed during cmpxchg. This is a ABA problem, for example, below case will happen:”…””}”(hj¬h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kahjºh²hubhŒtable”“”)”}”(hhh]”hŒtgroup”“”)”}”(hhh]”(hŒcolspec”“”)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K$uh1jÄhjÁubjÅ)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K#uh1jÄhjÁubhŒtbody”“”)”}”(hhh]”(hŒrow”“”)”}”(hhh]”hŒentry”“”)”}”(hhh]”(hï)”}”(hŒAt the beginning::”h]”hŒAt the beginning:”…””}”(hjéh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kfhjæubj)”}”(hŒvgpte = gfn1 gfn1 is mapped to pfn1 on host spte is the shadow page table entry corresponding with gpte and spte = pfn1”h]”hŒvgpte = gfn1 gfn1 is mapped to pfn1 on host spte is the shadow page table entry corresponding with gpte and spte = pfn1”…””}”hj÷sbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´Khhjæubeh}”(h]”h ]”h"]”h$]”h&]”Œmorecols”Kuh1jähjáubah}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubjà)”}”(hhh]”jå)”}”(hhh]”hï)”}”(hŒOn fast page fault path:”h]”hŒOn fast page fault path:”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kmhjubah}”(h]”h ]”h"]”h$]”h&]”Œmorecols”Kuh1jähjubah}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubjà)”}”(hhh]”(jå)”}”(hhh]”hï)”}”(hŒCPU 0:”h]”hŒCPU 0:”…””}”(hj9h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kohj6ubah}”(h]”h ]”h"]”h$]”h&]”uh1jähj3ubjå)”}”(hhh]”hï)”}”(hŒCPU 1:”h]”hŒCPU 1:”…””}”(hjPh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KohjMubah}”(h]”h ]”h"]”h$]”h&]”uh1jähj3ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubjà)”}”(hhh]”(jå)”}”(hhh]”j)”}”(hŒold_spte = *spte;”h]”hŒold_spte = *spte;”…””}”hjpsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´Kshjmubah}”(h]”h ]”h"]”h$]”h&]”uh1jähjjubjå)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”uh1jähjjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubjà)”}”(hhh]”(jå)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”uh1jähj“ubjå)”}”(hhh]”(hï)”}”(hŒpfn1 is swapped out::”h]”hŒpfn1 is swapped out:”…””}”(hj¢h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KuhjŸubj)”}”(hŒ spte = 0;”h]”hŒ spte = 0;”…””}”hj°sbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´KwhjŸubhï)”}”(hŒpfn1 is re-alloced for gfn2.”h]”hŒpfn1 is re-alloced for gfn2.”…””}”(hj¾h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KyhjŸubhï)”}”(hŒ/gpte is changed to point to gfn2 by the guest::”h]”hŒ.gpte is changed to point to gfn2 by the guest:”…””}”(hjÌh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K{hjŸubj)”}”(hŒspte = pfn1;”h]”hŒspte = pfn1;”…””}”hjÚsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K~hjŸubeh}”(h]”h ]”h"]”h$]”h&]”uh1jähj“ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubjà)”}”(hhh]”jå)”}”(hhh]”j)”}”(hŒ]if (cmpxchg(spte, old_spte, old_spte+W) mark_page_dirty(vcpu->kvm, gfn1) OOPS!!!”h]”hŒ]if (cmpxchg(spte, old_spte, old_spte+W) mark_page_dirty(vcpu->kvm, gfn1) OOPS!!!”…””}”hjúsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K‚hj÷ubah}”(h]”h ]”h"]”h$]”h&]”Œmorecols”Kuh1jähjôubah}”(h]”h ]”h"]”h$]”h&]”uh1jßhjÜubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÚhjÁubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1j¿hj¼ubah}”(h]”h ]”h"]”h$]”h&]”uh1jºhjºh²hh³hÇh´Nubhï)”}”(hŒ?We dirty-log for gfn1, that means gfn2 is lost in dirty-bitmap.”h]”hŒ?We dirty-log for gfn1, that means gfn2 is lost in dirty-bitmap.”…””}”(hj(h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K‡hjºh²hubhï)”}”(hŒFor direct sp, we can easily avoid it since the spte of direct sp is fixed to gfn. For indirect sp, we disabled fast page fault for simplicity.”h]”hŒFor direct sp, we can easily avoid it since the spte of direct sp is fixed to gfn. For indirect sp, we disabled fast page fault for simplicity.”…””}”(hj6h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K‰hjºh²hubhï)”}”(hŒZA solution for indirect sp could be to pin the gfn before the cmpxchg. After the pinning:”h]”hŒZA solution for indirect sp could be to pin the gfn before the cmpxchg. After the pinning:”…””}”(hjDh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KŒhjºh²hubhÿ)”}”(hhh]”(j)”}”(hŒdWe have held the refcount of pfn; that means the pfn can not be freed and be reused for another gfn.”h]”hï)”}”(hŒdWe have held the refcount of pfn; that means the pfn can not be freed and be reused for another gfn.”h]”hŒdWe have held the refcount of pfn; that means the pfn can not be freed and be reused for another gfn.”…””}”(hjYh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KhjUubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjRh²hh³hÇh´Nubj)”}”(hŒUThe pfn is writable and therefore it cannot be shared between different gfns by KSM. ”h]”hï)”}”(hŒTThe pfn is writable and therefore it cannot be shared between different gfns by KSM.”h]”hŒTThe pfn is writable and therefore it cannot be shared between different gfns by KSM.”…””}”(hjqh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K‘hjmubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjRh²hh³hÇh´Nubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´Khjºh²hubhï)”}”(hŒAThen, we can ensure the dirty bitmaps is correctly set for a gfn.”h]”hŒAThen, we can ensure the dirty bitmaps is correctly set for a gfn.”…””}”(hj‹h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K”hjºh²hubjè)”}”(hhh]”j)”}”(hŒDirty bit tracking ”h]”hï)”}”(hŒDirty bit tracking”h]”hŒDirty bit tracking”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K–hjœubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj™h²hh³hÇh´Nubah}”(h]”h ]”h"]”h$]”h&]”j"j#j$hj%j«Œstart”Kuh1jçhjºh²hh³hÇh´K–ubhï)”}”(hŒ¼In the original code, the spte can be fast updated (non-atomically) if the spte is read-only and the Accessed bit has already been set since the Accessed bit and Dirty bit can not be lost.”h]”hŒ¼In the original code, the spte can be fast updated (non-atomically) if the spte is read-only and the Accessed bit has already been set since the Accessed bit and Dirty bit can not be lost.”…””}”(hj»h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K˜hjºh²hubhï)”}”(hŒ‡But it is not true after fast page fault since the spte can be marked writable between reading spte and updating spte. Like below case:”h]”hŒ‡But it is not true after fast page fault since the spte can be marked writable between reading spte and updating spte. Like below case:”…””}”(hjÉh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kœhjºh²hubj»)”}”(hhh]”jÀ)”}”(hhh]”(jÅ)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K%uh1jÄhjÚubjÅ)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”Œcolwidth”K#uh1jÄhjÚubjÛ)”}”(hhh]”(jà)”}”(hhh]”jå)”}”(hhh]”(hï)”}”(hŒAt the beginning::”h]”hŒAt the beginning:”…””}”(hjúh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K hj÷ubj)”}”(hŒspte.W = 0 spte.Accessed = 1”h]”hŒspte.W = 0 spte.Accessed = 1”…””}”hjsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K¢hj÷ubeh}”(h]”h ]”h"]”h$]”h&]”Œmorecols”Kuh1jähjôubah}”(h]”h ]”h"]”h$]”h&]”uh1jßhjñubjà)”}”(hhh]”(jå)”}”(hhh]”hï)”}”(hŒCPU 0:”h]”hŒCPU 0:”…””}”(hj)h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K¥hj&ubah}”(h]”h ]”h"]”h$]”h&]”uh1jähj#ubjå)”}”(hhh]”hï)”}”(hŒCPU 1:”h]”hŒCPU 1:”…””}”(hj@h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K¥hj=ubah}”(h]”h ]”h"]”h$]”h&]”uh1jähj#ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjñubjà)”}”(hhh]”(jå)”}”(hhh]”(hï)”}”(hŒIn mmu_spte_update()::”h]”hŒIn mmu_spte_update():”…””}”(hj`h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K§hj]ubj)”}”(hŒ~old_spte = *spte; /* 'if' condition is satisfied. */ if (old_spte.Accessed == 1 && old_spte.W == 0) spte = new_spte;”h]”hŒ~old_spte = *spte; /* 'if' condition is satisfied. */ if (old_spte.Accessed == 1 && old_spte.W == 0) spte = new_spte;”…””}”hjnsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K©hj]ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jähjZubjå)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”uh1jähjZubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjñubjà)”}”(hhh]”(jå)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”uh1jähj‘ubjå)”}”(hhh]”(hï)”}”(hŒon fast page fault path::”h]”hŒon fast page fault path:”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K±hjubj)”}”(hŒ spte.W = 1”h]”hŒ spte.W = 1”…””}”hj®sbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K³hjubhï)”}”(hŒmemory write on the spte::”h]”hŒmemory write on the spte:”…””}”(hj¼h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kµhjubj)”}”(hŒspte.Dirty = 1”h]”hŒspte.Dirty = 1”…””}”hjÊsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K·hjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jähj‘ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjñubjà)”}”(hhh]”(jå)”}”(hhh]”j)”}”(hŒelse old_spte = xchg(spte, new_spte); if (old_spte.Accessed && !new_spte.Accessed) flush = true; if (old_spte.Dirty && !new_spte.Dirty) flush = true; OOPS!!!”h]”hŒelse old_spte = xchg(spte, new_spte); if (old_spte.Accessed && !new_spte.Accessed) flush = true; if (old_spte.Dirty && !new_spte.Dirty) flush = true; OOPS!!!”…””}”hjêsbah}”(h]”h ]”h"]”h$]”h&]”hÅhÆuh1jh³hÇh´K»hjçubah}”(h]”h ]”h"]”h$]”h&]”uh1jähjäubjå)”}”(hhh]”h}”(h]”h ]”h"]”h$]”h&]”uh1jähjäubeh}”(h]”h ]”h"]”h$]”h&]”uh1jßhjñubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÚhjÚubeh}”(h]”h ]”h"]”h$]”h&]”Œcols”Kuh1j¿hj×ubah}”(h]”h ]”h"]”h$]”h&]”uh1jºhjºh²hh³hÇh´Nubhï)”}”(hŒ#The Dirty bit is lost in this case.”h]”hŒ#The Dirty bit is lost in this case.”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÆhjºh²hubhï)”}”(hŒÐIn order to avoid this kind of issue, we always treat the spte as "volatile" if it can be updated out of mmu-lock [see spte_needs_atomic_update()]; it means the spte is always atomically updated in this case.”h]”hŒÔIn order to avoid this kind of issue, we always treat the spte as â€œvolatileâ€ if it can be updated out of mmu-lock [see spte_needs_atomic_update()]; it means the spte is always atomically updated in this case.”…””}”(hj.h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÈhjºh²hubjè)”}”(hhh]”j)”}”(hŒflush tlbs due to spte updated ”h]”hï)”}”(hŒflush tlbs due to spte updated”h]”hŒflush tlbs due to spte updated”…””}”(hjCh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÌhj?ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj<h²hh³hÇh´Nubah}”(h]”h ]”h"]”h$]”h&]”j"j#j$hj%j«jºKuh1jçhjºh²hh³hÇh´KÌubhï)”}”(hŒÃIf the spte is updated from writable to read-only, we should flush all TLBs, otherwise rmap_write_protect will find a read-only spte, even though the writable spte might be cached on a CPU's TLB.”h]”hŒÅIf the spte is updated from writable to read-only, we should flush all TLBs, otherwise rmap_write_protect will find a read-only spte, even though the writable spte might be cached on a CPUâ€™s TLB.”…””}”(hj]h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÎhjºh²hubhï)”}”(hXAs mentioned before, the spte can be updated to writable out of mmu-lock on fast page fault path. In order to easily audit the path, we see if TLBs needing to be flushed caused this reason in mmu_spte_update() since this is a common function to update spte (present -> present).”h]”hXAs mentioned before, the spte can be updated to writable out of mmu-lock on fast page fault path. In order to easily audit the path, we see if TLBs needing to be flushed caused this reason in mmu_spte_update() since this is a common function to update spte (present -> present).”…””}”(hjkh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÒhjºh²hubhï)”}”(hŒäSince the spte is "volatile" if it can be updated out of mmu-lock, we always atomically update the spte and the race caused by fast page fault can be avoided. See the comments in spte_needs_atomic_update() and mmu_spte_update().”h]”hŒèSince the spte is â€œvolatileâ€ if it can be updated out of mmu-lock, we always atomically update the spte and the race caused by fast page fault can be avoided. See the comments in spte_needs_atomic_update() and mmu_spte_update().”…””}”(hjyh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K×hjºh²hubhï)”}”(hŒLockless Access Tracking:”h]”hŒLockless Access Tracking:”…””}”(hj‡h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÛhjºh²hubhï)”}”(hX”This is used for Intel CPUs that are using EPT but do not support the EPT A/D bits. In this case, PTEs are tagged as A/D disabled (using ignored bits), and when the KVM MMU notifier is called to track accesses to a page (via kvm_mmu_notifier_clear_flush_young), it marks the PTE not-present in hardware by clearing the RWX bits in the PTE and storing the original R & X bits in more unused/ignored bits. When the VM tries to access the page later on, a fault is generated and the fast page fault mechanism described above is used to atomically restore the PTE to a Present state. The W bit is not saved when the PTE is marked for access tracking and during restoration to the Present state, the W bit is set depending on whether or not it was a write access. If it wasn't, then the W bit will remain clear until a write access happens, at which time it will be set using the Dirty tracking mechanism described above.”h]”hX–This is used for Intel CPUs that are using EPT but do not support the EPT A/D bits. In this case, PTEs are tagged as A/D disabled (using ignored bits), and when the KVM MMU notifier is called to track accesses to a page (via kvm_mmu_notifier_clear_flush_young), it marks the PTE not-present in hardware by clearing the RWX bits in the PTE and storing the original R & X bits in more unused/ignored bits. When the VM tries to access the page later on, a fault is generated and the fast page fault mechanism described above is used to atomically restore the PTE to a Present state. The W bit is not saved when the PTE is marked for access tracking and during restoration to the Present state, the W bit is set depending on whether or not it was a write access. If it wasnâ€™t, then the W bit will remain clear until a write access happens, at which time it will be set using the Dirty tracking mechanism described above.”…””}”(hj•h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KÝhjºh²hubeh}”(h]”Œ exception”ah ]”h"]”Œ2. exception”ah$]”h&]”uh1hÈhhÊh²hh³hÇh´KAubhÉ)”}”(hhh]”(hÎ)”}”(hŒ3. Reference”h]”hŒ3. Reference”…””}”(hj®h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhj«h²hh³hÇh´KëubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm_lock``”h]”jþ)”}”(hjÁh]”hŒkvm_lock”…””}”(hjÃh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhj¿ubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhj¼h²hh³hÇh´KîubhŒ field_list”“”)”}”(hhh]”(hŒfield”“”)”}”(hhh]”(hŒ field_name”“”)”}”(hŒType”h]”hŒType”…””}”(hjâh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjÝh³hÇh´KubhŒ field_body”“”)”}”(hŒmutex”h]”hï)”}”(hjôh]”hŒmutex”…””}”(hjöh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kðhjòubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjÝubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´KðhjØh²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjh³hÇh´Kubjñ)”}”(hŒany”h]”hï)”}”(hj"h]”hŒany”…””}”(hj$h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kñhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´KñhjØh²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hj@h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj=h³hÇh´Kubjñ)”}”(hŒ - vm_list ”h]”hÿ)”}”(hhh]”j)”}”(hŒvm_list ”h]”hï)”}”(hŒvm_list”h]”hŒvm_list”…””}”(hjYh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´KòhjUubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjRubah}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´KòhjNubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj=ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´KòhjØh²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhj¼h²hh³hÇh´Kðubeh}”(h]”Œkvm-lock”ah ]”h"]”Œkvm_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´KîubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm_usage_lock``”h]”jþ)”}”(hj’h]”hŒkvm_usage_lock”…””}”(hj”h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjh²hh³hÇh´Kõubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjªh³hÇh´Kubjñ)”}”(hŒmutex”h]”hï)”}”(hj½h]”hŒmutex”…””}”(hj¿h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´K÷hj»ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjªubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´K÷hj§h²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjÛh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjØh³hÇh´Kubjñ)”}”(hŒany”h]”hï)”}”(hjëh]”hŒany”…””}”(hjíh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Køhjéubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjØubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Køhj§h²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj h³hÇh´Kubjñ)”}”(hŒ:- kvm_usage_count - hardware virtualization enable/disable”h]”hÿ)”}”(hhh]”(j)”}”(hŒkvm_usage_count”h]”hï)”}”(hj h]”hŒkvm_usage_count”…””}”(hj" h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kùhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj ubj)”}”(hŒ&hardware virtualization enable/disable”h]”hï)”}”(hj7 h]”hŒ&hardware virtualization enable/disable”…””}”(hj9 h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kúhj5 ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhj ubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´Kùhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Kùhj§h²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hja h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj^ h³hÇh´Kubjñ)”}”(hŒ€Exists to allow taking cpus_read_lock() while kvm_usage_count is protected, which simplifies the virtualization enabling logic. ”h]”hï)”}”(hŒExists to allow taking cpus_read_lock() while kvm_usage_count is protected, which simplifies the virtualization enabling logic.”h]”hŒExists to allow taking cpus_read_lock() while kvm_usage_count is protected, which simplifies the virtualization enabling logic.”…””}”(hjs h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Kûhjo ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj^ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Kûhj§h²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhjh²hh³hÇh´K÷ubeh}”(h]”Œkvm-usage-lock”ah ]”h"]”Œkvm_usage_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´KõubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm->mn_invalidate_lock``”h]”jþ)”}”(hj h]”hŒkvm->mn_invalidate_lock”…””}”(hj¢ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjž ubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhj› h²hh³hÇh´Kÿubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hj» h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj¸ h³hÇh´Kubjñ)”}”(hŒ spinlock_t”h]”hï)”}”(hjË h]”hŒ spinlock_t”…””}”(hjÍ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´MhjÉ ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj¸ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjµ h²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjé h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjæ h³hÇh´Kubjñ)”}”(hŒany”h]”hï)”}”(hjù h]”hŒany”…””}”(hjû h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj÷ ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjæ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjµ h²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj h³hÇh´Kubjñ)”}”(hŒ7mn_active_invalidate_count, mn_memslots_update_rcuwait ”h]”hï)”}”(hŒ6mn_active_invalidate_count, mn_memslots_update_rcuwait”h]”hŒ6mn_active_invalidate_count, mn_memslots_update_rcuwait”…””}”(hj) h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj% ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjµ h²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhj› h²hh³hÇh´Mubeh}”(h]”Œkvm-mn-invalidate-lock”ah ]”h"]”Œkvm->mn_invalidate_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´KÿubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm_arch::tsc_write_lock``”h]”jþ)”}”(hjV h]”hŒkvm_arch::tsc_write_lock”…””}”(hjX h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjT ubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjQ h²hh³hÇh´Mubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjq h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjn h³hÇh´Kubjñ)”}”(hŒraw_spinlock_t”h]”hï)”}”(hj h]”hŒraw_spinlock_t”…””}”(hjƒ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjn ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjŸ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjœ h³hÇh´Kubjñ)”}”(hŒx86”h]”hï)”}”(hj¯ h]”hŒx86”…””}”(hj± h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M hj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjœ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M hjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hjÍ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjÊ h³hÇh´Kubjñ)”}”(hŒO- kvm_arch::{last_tsc_write,last_tsc_nsec,last_tsc_offset} - tsc offset in vmcb”h]”hÿ)”}”(hhh]”(j)”}”(hŒ8kvm_arch::{last_tsc_write,last_tsc_nsec,last_tsc_offset}”h]”hï)”}”(hjä h]”hŒ8kvm_arch::{last_tsc_write,last_tsc_nsec,last_tsc_offset}”…””}”(hjæ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M hjâ ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjß ubj)”}”(hŒtsc offset in vmcb”h]”hï)”}”(hjû h]”hŒtsc offset in vmcb”…””}”(hjý h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjù ubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjß ubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´M hjÛ ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjÊ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M hjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hj%h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj"h³hÇh´Kubjñ)”}”(hŒ>'raw' because updating the tsc offsets must not be preempted. ”h]”hï)”}”(hŒ='raw' because updating the tsc offsets must not be preempted.”h]”hŒAâ€˜rawâ€™ because updating the tsc offsets must not be preempted.”…””}”(hj7h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj3ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj"ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjk h²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhjQ h²hh³hÇh´Mubeh}”(h]”Œkvm-arch-tsc-write-lock”ah ]”h"]”Œkvm_arch::tsc_write_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´MubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm->mmu_lock``”h]”jþ)”}”(hjdh]”hŒ kvm->mmu_lock”…””}”(hjfh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjbubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhj_h²hh³hÇh´Mubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj|h³hÇh´Kubjñ)”}”(hŒspinlock_t or rwlock_t”h]”hï)”}”(hjh]”hŒspinlock_t or rwlock_t”…””}”(hj‘h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj|ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjyh²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjªh³hÇh´Kubjñ)”}”(hŒany”h]”hï)”}”(hj½h]”hŒany”…””}”(hj¿h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj»ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjªubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjyh²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hjÛh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjØh³hÇh´Kubjñ)”}”(hŒ-shadow page/shadow tlb entry”h]”hï)”}”(hjëh]”hŒ-shadow page/shadow tlb entry”…””}”(hjíh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjéubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjØubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjyh²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjh³hÇh´Kubjñ)”}”(hŒ3it is a spinlock since it is used in mmu notifier. ”h]”hï)”}”(hŒ2it is a spinlock since it is used in mmu notifier.”h]”hŒ2it is a spinlock since it is used in mmu notifier.”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhjyh²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhj_h²hh³hÇh´Mubeh}”(h]”Œkvm-mmu-lock”ah ]”h"]”Œ kvm->mmu_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´MubhÉ)”}”(hhh]”(hÎ)”}”(hŒ ``kvm->srcu``”h]”jþ)”}”(hjHh]”hŒ kvm->srcu”…””}”(hjJh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjFubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjCh²hh³hÇh´Mubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjch²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj`h³hÇh´Kubjñ)”}”(hŒ srcu lock”h]”hï)”}”(hjsh]”hŒ srcu lock”…””}”(hjuh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjqubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj`ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhj]h²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hj‘h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjŽh³hÇh´Kubjñ)”}”(hŒany”h]”hï)”}”(hj¡h]”hŒany”…””}”(hj£h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´MhjŸubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjŽubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhj]h²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hj¿h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj¼h³hÇh´Kubjñ)”}”(hŒ- kvm->memslots - kvm->buses”h]”hÿ)”}”(hhh]”(j)”}”(hŒ kvm->memslots”h]”hï)”}”(hjÖh]”hŒ kvm->memslots”…””}”(hjØh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´MhjÔubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÑubj)”}”(hŒ kvm->buses”h]”hï)”}”(hjíh]”hŒ kvm->buses”…””}”(hjïh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhjëubah}”(h]”h ]”h"]”h$]”h&]”uh1jhjÑubeh}”(h]”h ]”h"]”h$]”h&]”j³j´uh1hþh³hÇh´MhjÍubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj¼ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhj]h²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj h³hÇh´Kubjñ)”}”(hXThe srcu read lock must be held while accessing memslots (e.g. when using gfn_to_* functions) and while accessing in-kernel MMIO/PIO address->device structure mapping (kvm->buses). The srcu index can be stored in kvm_vcpu->srcu_idx per vcpu if it is needed by multiple functions. ”h]”hï)”}”(hXThe srcu read lock must be held while accessing memslots (e.g. when using gfn_to_* functions) and while accessing in-kernel MMIO/PIO address->device structure mapping (kvm->buses). The srcu index can be stored in kvm_vcpu->srcu_idx per vcpu if it is needed by multiple functions.”h]”hXThe srcu read lock must be held while accessing memslots (e.g. when using gfn_to_* functions) and while accessing in-kernel MMIO/PIO address->device structure mapping (kvm->buses). The srcu index can be stored in kvm_vcpu->srcu_idx per vcpu if it is needed by multiple functions.”…””}”(hj) h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´Mhj% ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´Mhj]h²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhjCh²hh³hÇh´Mubeh}”(h]”Œkvm-srcu”ah ]”h"]”Œ kvm->srcu”ah$]”h&]”uh1hÈhj«h²hh³hÇh´MubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``kvm->slots_arch_lock``”h]”jþ)”}”(hjV h]”hŒkvm->slots_arch_lock”…””}”(hjX h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjT ubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjQ h²hh³hÇh´M"ubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjq h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjn h³hÇh´Kubjñ)”}”(hŒmutex”h]”hï)”}”(hj h]”hŒmutex”…””}”(hjƒ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M#hj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjn ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M#hjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjŸ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjœ h³hÇh´Kubjñ)”}”(hŒany (only needed on x86 though)”h]”hï)”}”(hj¯ h]”hŒany (only needed on x86 though)”…””}”(hj± h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M$hj ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjœ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M$hjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hjÍ h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjÊ h³hÇh´Kubjñ)”}”(hŒlany arch-specific fields of memslots that have to be modified in a ``kvm->srcu`` read-side critical section.”h]”hï)”}”(hŒlany arch-specific fields of memslots that have to be modified in a ``kvm->srcu`` read-side critical section.”h]”(hŒCany arch-specific fields of memslots that have to be modified in a ”…””}”(hjß h²hh³Nh´Nubjþ)”}”(hŒ ``kvm->srcu``”h]”hŒ kvm->srcu”…””}”(hjç h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjß ubhŒ read-side critical section.”…””}”(hjß h²hh³Nh´Nubeh}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M%hjÛ ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjÊ ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M%hjk h²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjh³hÇh´Kubjñ)”}”(hŒvmust be held before reading the pointer to the current memslots, until after all changes to the memslots are complete ”h]”hï)”}”(hŒumust be held before reading the pointer to the current memslots, until after all changes to the memslots are complete”h]”hŒumust be held before reading the pointer to the current memslots, until after all changes to the memslots are complete”…””}”(hj h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M'hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M'hjk h²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhjQ h²hh³hÇh´M#ubeh}”(h]”Œkvm-slots-arch-lock”ah ]”h"]”Œkvm->slots_arch_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´M"ubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``wakeup_vcpus_on_cpu_lock``”h]”jþ)”}”(hjMh]”hŒwakeup_vcpus_on_cpu_lock”…””}”(hjOh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhjKubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhjHh²hh³hÇh´M+ubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjhh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjeh³hÇh´Kubjñ)”}”(hŒ spinlock_t”h]”hï)”}”(hjxh]”hŒ spinlock_t”…””}”(hjzh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M,hjvubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjeubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M,hjbh²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hj–h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj“h³hÇh´Kubjñ)”}”(hŒx86”h]”hï)”}”(hj¦h]”hŒx86”…””}”(hj¨h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M-hj¤ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj“ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M-hjbh²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hjÄh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjÁh³hÇh´Kubjñ)”}”(hŒwakeup_vcpus_on_cpu”h]”hï)”}”(hjÔh]”hŒwakeup_vcpus_on_cpu”…””}”(hjÖh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M.hjÒubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjÁubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M.hjbh²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hjòh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjïh³hÇh´Kubjñ)”}”(hXŠThis is a per-CPU lock and it is used for VT-d posted-interrupts. When VT-d posted-interrupts are supported and the VM has assigned devices, we put the blocked vCPU on the list blocked_vcpu_on_cpu protected by blocked_vcpu_on_cpu_lock. When VT-d hardware issues wakeup notification event since external interrupts from the assigned devices happens, we will find the vCPU on the list to wakeup. ”h]”hï)”}”(hX‰This is a per-CPU lock and it is used for VT-d posted-interrupts. When VT-d posted-interrupts are supported and the VM has assigned devices, we put the blocked vCPU on the list blocked_vcpu_on_cpu protected by blocked_vcpu_on_cpu_lock. When VT-d hardware issues wakeup notification event since external interrupts from the assigned devices happens, we will find the vCPU on the list to wakeup.”h]”hX‰This is a per-CPU lock and it is used for VT-d posted-interrupts. When VT-d posted-interrupts are supported and the VM has assigned devices, we put the blocked vCPU on the list blocked_vcpu_on_cpu protected by blocked_vcpu_on_cpu_lock. When VT-d hardware issues wakeup notification event since external interrupts from the assigned devices happens, we will find the vCPU on the list to wakeup.”…””}”(hjh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M/hjubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjïubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M/hjbh²hubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÖhjHh²hh³hÇh´M,ubeh}”(h]”Œwakeup-vcpus-on-cpu-lock”ah ]”h"]”Œwakeup_vcpus_on_cpu_lock”ah$]”h&]”uh1hÈhj«h²hh³hÇh´M+ubhÉ)”}”(hhh]”(hÎ)”}”(hŒ``vendor_module_lock``”h]”jþ)”}”(hj1h]”hŒvendor_module_lock”…””}”(hj3h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jýhj/ubah}”(h]”h ]”h"]”h$]”h&]”uh1hÍhj,h²hh³hÇh´M8ubj×)”}”(hhh]”(jÜ)”}”(hhh]”(já)”}”(hŒType”h]”hŒType”…””}”(hjLh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjIh³hÇh´Kubjñ)”}”(hŒmutex”h]”hï)”}”(hj\h]”hŒmutex”…””}”(hj^h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M9hjZubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjIubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M9hjFh²hubjÜ)”}”(hhh]”(já)”}”(hŒArch”h]”hŒArch”…””}”(hjzh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjwh³hÇh´Kubjñ)”}”(hŒx86”h]”hï)”}”(hjŠh]”hŒx86”…””}”(hjŒh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M:hjˆubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhjwubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M:hjFh²hubjÜ)”}”(hhh]”(já)”}”(hŒProtects”h]”hŒProtects”…””}”(hj¨h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhj¥h³hÇh´Kubjñ)”}”(hŒ.loading a vendor module (kvm_amd or kvm_intel)”h]”hï)”}”(hj¸h]”hŒ.loading a vendor module (kvm_amd or kvm_intel)”…””}”(hjºh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´M;hj¶ubah}”(h]”h ]”h"]”h$]”h&]”uh1jðhj¥ubeh}”(h]”h ]”h"]”h$]”h&]”uh1jÛh³hÇh´M;hjFh²hubjÜ)”}”(hhh]”(já)”}”(hŒComment”h]”hŒComment”…””}”(hjÖh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1jàhjÓh³hÇh´Kubjñ)”}”(hX>Exists because using kvm_lock leads to deadlock. kvm_lock is taken in notifiers, e.g. __kvmclock_cpufreq_notifier(), that may be invoked while cpu_hotplug_lock is held, e.g. from cpufreq_boost_trigger_state(), and many operations need to take cpu_hotplug_lock when loading a vendor module, e.g. updating static calls.”h]”hï)”}”(hX>Exists because using kvm_lock leads to deadlock. kvm_lock is taken in notifiers, e.g. __kvmclock_cpufreq_notifier(), that may be invoked while cpu_hotplug_lock is held, e.g. from cpufreq_boost_trigger_state(), and many operations need to take cpu_hotplug_lock when loading a vendor module, e.g. updating static calls.”h]”hX>Exists because using kvm_lock leads to deadlock. kvm_lock is taken in notifiers, e.g. __kvmclock_cpufreq_notifier(), that may be invoked while cpu_hotplug_lock is held, e.g. from cpufreq_boost_trigger_state(), and many operations need to take cpu_hotplug_lock when loading a vendor module, e.g. updating static calls.”…””}”(hjèh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîh³hÇh´MEnumerated list start value not ordinal-1: â€œ2â€ (ordinal 2)”…””}”(hj©h²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîhj¦ubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”ŒINFO”Œsource”hÇŒline”Kuh1j¤hjºh²hh³hÇh´K–ubj¥)”}”(hhh]”hï)”}”(hŒ:Enumerated list start value not ordinal-1: "3" (ordinal 3)”h]”hŒ>Enumerated list start value not ordinal-1: â€œ3â€ (ordinal 3)”…””}”(hjÅh²hh³Nh´Nubah}”(h]”h ]”h"]”h$]”h&]”uh1hîhjÂubah}”(h]”h ]”h"]”h$]”h&]”Œlevel”KŒtype”j¿Œsource”hÇŒline”Kuh1j¤hjºh²hh³hÇh´KÌubeŒtransform_messages”]”Œtransformer”NŒinclude_log”]”Œ decoration”Nh²hub.