- 08 Dec, 2021 6 commits
-
-
Marc Zyngier authored
Everywhere we use kvm_for_each_vpcu(), we use an int as the vcpu index. Unfortunately, we're about to move rework the iterator, which requires this to be upgrade to an unsigned long. Let's bite the bullet and repaint all of it in one go. Signed-off-by:
Marc Zyngier <maz@kernel.org> Message-Id: <20211116160403.4074052-7-maz@kernel.org> Signed-off-by:
Paolo Bonzini <pbonzini@redhat.com>
-
Marc Zyngier authored
At least on arm64 and x86, the vcpus array is pretty huge (up to 1024 entries on x86) and is mostly empty in the majority of the cases (running 1k vcpu VMs is not that common). This mean that we end-up with a 4kB block of unused memory in the middle of the kvm structure. Instead of wasting away this memory, let's use an xarray instead, which gives us almost the same flexibility as a normal array, but with a reduced memory usage with smaller VMs. Signed-off-by:
-
Marc Zyngier authored
As we are about to change the way vcpus are allocated, mandate the use of kvm_get_vcpu() instead of open-coding the access. Reviewed-by:
Claudio Imbrenda <imbrenda@linux.ibm.com> Signed-off-by:
-
Marc Zyngier authored
As we are about to change the way vcpus are allocated, mandate the use of kvm_get_vcpu() instead of open-coding the access. Reviewed-by:
Philippe Mathieu-Daudé <f4bug@amsat.org> Signed-off-by:
-
Marc Zyngier authored
All architectures have similar loops iterating over the vcpus, freeing one vcpu at a time, and eventually wiping the reference off the vcpus array. They are also inconsistently taking the kvm->lock mutex when wiping the references from the array. Make this code common, which will simplify further changes. The locking is dropped altogether, as this should only be called when there is no further references on the kvm structure. Reviewed-by:
-
Paolo Bonzini authored
Fix the number of bits in the role, and simplify the explanation of why several bits or combinations of bits are redundant. Signed-off-by:
-
- 05 Dec, 2021 3 commits
-
-
Tom Lendacky authored
Currently, an SEV-ES guest is terminated if the validation of the VMGEXIT exit code or exit parameters fails. The VMGEXIT instruction can be issued from userspace, even though userspace (likely) can't update the GHCB. To prevent userspace from being able to kill the guest, return an error through the GHCB when validation fails rather than terminating the guest. For cases where the GHCB can't be updated (e.g. the GHCB can't be mapped, etc.), just return back to the guest. The new error codes are documented in the lasest update to the GHCB specification. Fixes: 291bd20d ("KVM: SVM: Add initial support for a VMGEXIT VMEXIT") Signed-off-by:
Tom Lendacky <thomas.lendacky@amd.com> Message-Id: <b57280b5562893e2616257ac9c2d4525a9aeeb42.1638471124.git.thomas.lendacky@amd.com> Signed-off-by:
-
Sean Christopherson authored
Use kvzalloc() to allocate KVM's buffer for SEV-ES's GHCB scratch area so that KVM falls back to __vmalloc() if physically contiguous memory isn't available. The buffer is purely a KVM software construct, i.e. there's no need for it to be physically contiguous. Cc: Tom Lendacky <thomas.lendacky@amd.com> Signed-off-by:
Sean Christopherson <seanjc@google.com> Message-Id: <20211109222350.2266045-3-seanjc@google.com> Signed-off-by:
-
Sean Christopherson authored
Return appropriate error codes if setting up the GHCB scratch area for an SEV-ES guest fails. In particular, returning -EINVAL instead of -ENOMEM when allocating the kernel buffer could be confusing as userspace would likely suspect a guest issue. Fixes: 8f423a80 ("KVM: SVM: Support MMIO for an SEV-ES guest") Cc: Tom Lendacky <thomas.lendacky@amd.com> Signed-off-by:
-
- 02 Dec, 2021 4 commits
-
-
Sean Christopherson authored
Bail from the page fault handler if the root shadow page was obsoleted by a memslot update. Do the check _after_ acuiring mmu_lock, as the TDP MMU doesn't rely on the memslot/MMU generation, and instead relies on the root being explicit marked invalid by kvm_mmu_zap_all_fast(), which takes mmu_lock for write. For the TDP MMU, inserting a SPTE into an obsolete root can leak a SP if kvm_tdp_mmu_zap_invalidated_roots() has already zapped the SP, i.e. has moved past the gfn associated with the SP. For other MMUs, the resulting behavior is far more convoluted, though unlikely to be truly problematic. Installing SPs/SPTEs into the obsolete root isn't directly problematic, as the obsolete root will be unloaded and dropped before the vCPU re-enters the guest. But because the legacy MMU tracks shadow pages by their role, any SP created by the fault can can be reused in the new post-reload root. Again, that _shouldn't_ be problematic as any leaf child SPTEs will be created for the current/valid memslot generation, and kvm_mmu_get_page() will not reuse child SPs from the old generation as they will be flagged as obsolete. But, given that continuing with the fault is pointess (the root will be unloaded), apply the check to all MMUs. Fixes: b7cccd39 ("KVM: x86/mmu: Fast invalidation for TDP MMU") Cc: stable@vger.kernel.org Cc: Ben Gardon <bgardon@google.com> Signed-off-by:
-
Dan Carpenter authored
The error paths in the prepare_vmcs02() function are supposed to set *entry_failure_code but this path does not. It leads to using an uninitialized variable in the caller. Fixes: 71f73470 ("KVM: nVMX: Load GUEST_IA32_PERF_GLOBAL_CTRL MSR on VM-Entry") Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Message-Id: <20211130125337.GB24578@kili> Signed-off-by:
-
Paolo Bonzini authored
kvm_vcpu_apicv_active() returns false if a virtual machine has no in-kernel local APIC, however kvm_apicv_activated might still be true if there are no reasons to disable APICv; in fact it is quite likely that there is none because APICv is inhibited by specific configurations of the local APIC and those configurations cannot be programmed. This triggers a WARN: WARN_ON_ONCE(kvm_apicv_activated(vcpu->kvm) != kvm_vcpu_apicv_active(vcpu)); To avoid this, introduce another cause for APICv inhibition, namely the absence of an in-kernel local APIC. This cause is enabled by default, and is dropped by either KVM_CREATE_IRQCHIP or the enabling of KVM_CAP_IRQCHIP_SPLIT. Reported-by:
Ignat Korchagin <ignat@cloudflare.com> Fixes: ee49a893 ("KVM: x86: Move SVM's APICv sanity check to common x86", 2021-10-22) Reviewed-by:
Maxim Levitsky <mlevitsk@redhat.com> Reviewed-by:
-
Like Xu authored
If we run the following perf command in an AMD Milan guest: perf stat \ -e cpu/event=0x1d0/ \ -e cpu/event=0x1c7/ \ -e cpu/umask=0x1f,event=0x18e/ \ -e cpu/umask=0x7,event=0x18e/ \ -e cpu/umask=0x18,event=0x18e/ \ ./workload dmesg will report a #GP warning from an unchecked MSR access error on MSR_F15H_PERF_CTLx. This is because according to APM (Revision: 4.03) Figure 13-7, the bits [35:32] of AMD PerfEvtSeln register is a part of the event select encoding, which extends the EVENT_SELECT field from 8 bits to 12 bits. Opportunistically update pmu->reserved_bits for reserved bit 19. Reported-by:
Jim Mattson <jmattson@google.com> Fixes: ca724305 ("KVM: x86/vPMU: Implement AMD vPMU code for KVM") Signed-off-by:
Like Xu <likexu@tencent.com> Message-Id: <20211118130320.95997-1-likexu@tencent.com> Signed-off-by:
-
- 30 Nov, 2021 19 commits
-
-
Paolo Bonzini authored
avic_set_running() passes the current CPU to avic_vcpu_load(), albeit via vcpu->cpu rather than smp_processor_id(). If the thread is migrated while avic_set_running runs, the call to avic_vcpu_load() can use a stale value for the processor id. Avoid this by blocking preemption over the entire execution of avic_set_running(). Reported-by:
-
Paolo Bonzini authored
There is nothing to synchronize if APICv is disabled, since neither other vCPUs nor assigned devices can set PIR.ON. Signed-off-by: -
Paolo Bonzini authored
Generally, kvm->lock is not taken for a long time, but sev_lock_two_vms is different: it takes vCPU locks inside, so userspace can hold it back just by calling a vCPU ioctl. Play it safe and use mutex_lock_killable. Message-Id: <20211123005036.2954379-13-pbonzini@redhat.com> Signed-off-by: -
Paolo Bonzini authored
Taking the lock is useless since there are no other references, and there are already accesses (e.g. to sev->enc_context_owner) that do not take it. So get rid of it. Reviewed-by:
-
Paolo Bonzini authored
VMs that mirror an encryption context rely on the owner to keep the ASID allocated. Performing a KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM would cause a dangling ASID: 1. copy context from A to B (gets ref to A) 2. move context from A to L (moves ASID from A to L) 3. close L (releases ASID from L, B still references it) The right way to do the handoff instead is to create a fresh mirror VM on the destination first: 1. copy context from A to B (gets ref to A) [later] 2. close B (releases ref to A) 3. move context from A to L (moves ASID from A to L) 4. copy context from L to M So, catch the situation by adding a count of how many VMs are mirroring this one's encryption context. Fixes: 0b020f5a ("KVM: SEV: Add support for SEV-ES intra host migration") Message-Id: <20211123005036.2954379-11-pbonzini@redhat.com> Signed-off-by:
-
Paolo Bonzini authored
Now that we have a facility to lock two VMs with deadlock protection, use it for the creation of mirror VMs as well. One of COPY_ENC_CONTEXT_FROM(dst, src) and COPY_ENC_CONTEXT_FROM(src, dst) would always fail, so the combination is nonsensical and it is okay to return -EBUSY if it is attempted. This sidesteps the question of what happens if a VM is MOVE_ENC_CONTEXT_FROM'd at the same time as it is COPY_ENC_CONTEXT_FROM'd: the locking prevents that from happening. Cc: Peter Gonda <pgonda@google.com> Cc: Sean Christopherson <seanjc@google.com> Reviewed-by:
-
Paolo Bonzini authored
I am putting the tests in sev_migrate_tests because the failure conditions are very similar and some of the setup code can be reused, too. The tests cover both successful creation of a mirror VM, and error conditions. Cc: Peter Gonda <pgonda@google.com> Cc: Sean Christopherson <seanjc@google.com> Message-Id: <20211123005036.2954379-9-pbonzini@redhat.com> Signed-off-by: -
Paolo Bonzini authored
Allow intra-host migration of a mirror VM; the destination VM will be a mirror of the same ASID as the source. Fixes: b5663931 ("KVM: SEV: Add support for SEV intra host migration") Reviewed-by:
-
Paolo Bonzini authored
This was broken before the introduction of KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM, but technically harmless because the region list was unused for a mirror VM. However, it is untidy and it now causes a NULL pointer access when attempting to move the encryption context of a mirror VM. Fixes: 54526d1f ("KVM: x86: Support KVM VMs sharing SEV context") Message-Id: <20211123005036.2954379-7-pbonzini@redhat.com> Signed-off-by:
-
Paolo Bonzini authored
Encapsulate the handling of the migration_in_progress flag for both VMs in two functions sev_lock_two_vms and sev_unlock_two_vms. It does not matter if KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM locks the destination struct kvm a bit later, and this change 1) keeps the cleanup chain of labels smaller 2) makes it possible for KVM_CAP_VM_COPY_ENC_CONTEXT_FROM to reuse the logic. Cc: Peter Gonda <pgonda@google.com> Cc: Sean Christopherson <seanjc@google.com> Message-Id: <20211123005036.2954379-6-pbonzini@redhat.com> Signed-off-by: -
Paolo Bonzini authored
list_replace_init cannot be used if the source is an empty list, because "new->next->prev = new" will overwrite "old->next": new old prev = new, next = new prev = old, next = old new->next = old->next prev = new, next = old prev = old, next = old new->next->prev = new prev = new, next = old prev = old, next = new new->prev = old->prev prev = old, next = old prev = old, next = old new->next->prev = new prev = old, next = old prev = new, next = new The desired outcome instead would be to leave both old and new the same as they were (two empty circular lists). Use list_cut_before, which already has the necessary check and is documented to discard the previous contents of the list that will hold the result. Fixes: b5663931 ("KVM: SEV: Add support for SEV intra host migration") Reviewed-by:
-
Paolo Bonzini authored
Currently, checks for whether VT-d PI can be used refer to the current status of the feature in the current vCPU; or they more or less pick vCPU 0 in case a specific vCPU is not available. However, these checks do not attempt to synchronize with changes to the IRTE. In particular, there is no path that updates the IRTE when APICv is re-activated on vCPU 0; and there is no path to wakeup a CPU that has APICv disabled, if the wakeup occurs because of an IRTE that points to a posted interrupt. To fix this, always go through the VT-d PI path as long as there are assigned devices and APICv is available on both the host and the VM side. Since the relevant condition was copied over three times, take the hint and factor it into a separate function. Suggested-by:
David Matlack <dmatlack@google.com> Message-Id: <20211123004311.2954158-5-pbonzini@redhat.com> Signed-off-by:
-
Paolo Bonzini authored
The IRTE for an assigned device can trigger a POSTED_INTR_VECTOR even if APICv is disabled on the vCPU that receives it. In that case, the interrupt will just cause a vmexit and leave the ON bit set together with the PIR bit corresponding to the interrupt. Right now, the interrupt would not be delivered until APICv is re-enabled. However, fixing this is just a matter of always doing the PIR->IRR synchronization, even if the vCPU has temporarily disabled APICv. This is not a problem for performance, or if anything it is an improvement. First, in the common case where vcpu->arch.apicv_active is true, one fewer check has to be performed. Second, static_call_cond will elide the function call if APICv is not present or disabled. Finally, in the case for AMD hardware we can remove the sync_pir_to_irr callback: it is only needed for apic_has_interrupt_for_ppr, and that function already has a fallback for !APICv. Cc: stable@vger.kernel.org Co-developed-by:
-
Paolo Bonzini authored
If APICv is disabled for this vCPU, assigned devices may still attempt to post interrupts. In that case, we need to cancel the vmentry and deliver the interrupt with KVM_REQ_EVENT. Extend the existing code that handles injection of L1 interrupts into L2 to cover this case as well. vmx_hwapic_irr_update is only called when APICv is active so it would be confusing to add a check for vcpu->arch.apicv_active in there. Instead, just use vmx_set_rvi directly in vmx_sync_pir_to_irr. Cc: stable@vger.kernel.org Reviewed-by:
-
Maciej S. Szmigiero authored
A kvm_page_table_test run with its default settings fails on VMX due to memory region add failure: > ==== Test Assertion Failure ==== > lib/kvm_util.c:952: ret == 0 > pid=10538 tid=10538 errno=17 - File exists > 1 0x00000000004057d1: vm_userspace_mem_region_add at kvm_util.c:947 > 2 0x0000000000401ee9: pre_init_before_test at kvm_page_table_test.c:302 > 3 (inlined by) run_test at kvm_page_table_test.c:374 > 4 0x0000000000409754: for_each_guest_mode at guest_modes.c:53 > 5 0x0000000000401860: main at kvm_page_table_test.c:500 > 6 0x00007f82ae2d8554: ?? ??:0 > 7 0x0000000000401894: _start at ??:? > KVM_SET_USER_MEMORY_REGION IOCTL failed, > rc: -1 errno: 17 > slot: 1 flags: 0x0 > guest_phys_addr: 0xc0000000 size: 0x40000000 This is because the memory range that this test is trying to add (0x0c0000000 - 0x100000000) conflicts with LAPIC mapping at 0x0fee00000. Looking at the code it seems that guest_test_*phys*_mem variable gets mistakenly overwritten with guest_test_*virt*_mem while trying to adjust the former for alignment. With the correct variable adjusted this test runs successfully. Signed-off-by:
Maciej S. Szmigiero <maciej.szmigiero@oracle.com> Message-Id: <52e487458c3172923549bbcf9dfccfbe6faea60b.1637940473.git.maciej.szmigiero@oracle.com> Signed-off-by:
-
Sean Christopherson authored
Account for the '0' being a default, "let KVM choose" period, when determining whether or not the recovery worker needs to be awakened in response to userspace reducing the period. Failure to do so results in the worker not being awakened properly, e.g. when changing the period from '0' to any small-ish value. Fixes: 4dfe4f40 ("kvm: x86: mmu: Make NX huge page recovery period configurable") Cc: stable@vger.kernel.org Cc: Junaid Shahid <junaids@google.com> Signed-off-by:
-
Paolo Bonzini authored
Initialize the mask for PKU permissions as if CR4.PKE=0, avoiding incorrect interpretations of the nested hypervisor's page tables. Cc: stable@vger.kernel.org Signed-off-by: -
Sean Christopherson authored
Drop the "flush" param and return values to/from the TDP MMU's helper for zapping collapsible SPTEs. Because the helper runs with mmu_lock held for read, not write, it uses tdp_mmu_zap_spte_atomic(), and the atomic zap handles the necessary remote TLB flush. Similarly, because mmu_lock is dropped and re-acquired between zapping legacy MMUs and zapping TDP MMUs, kvm_mmu_zap_collapsible_sptes() must handle remote TLB flushes from the legacy MMU before calling into the TDP MMU. Fixes: e2209710 ("KVM: x86/mmu: Skip rmap operations if rmaps not allocated") Signed-off-by:
-
Sean Christopherson authored
Use the yield-safe variant of the TDP MMU iterator when handling an unmapping event from the MMU notifier, as most occurences of the event allow yielding. Fixes: e1eed584 ("KVM: x86/mmu: Allow yielding during MMU notifier unmap/zap, if possible") Cc: stable@vger.kernel.org Signed-off-by:
-
- 26 Nov, 2021 8 commits
-
-
Lai Jiangshan authored
INVLPG operates on guest virtual address, which are represented by vcpu->arch.walk_mmu. In nested virtualization scenarios, kvm_mmu_invlpg() was using the wrong MMU structure; if L2's invlpg were emulated by L0 (in practice, it hardly happen) when nested two-dimensional paging is enabled, the call to ->tlb_flush_gva() would be skipped and the hardware TLB entry would not be invalidated. Signed-off-by:
Lai Jiangshan <laijs@linux.alibaba.com> Message-Id: <20211124122055.64424-5-jiangshanlai@gmail.com> Cc: stable@vger.kernel.org Signed-off-by:
-
Lai Jiangshan authored
If the is an L1 with nNPT in 32bit, the shadow walk starts with pae_root. Fixes: a717a780 ("KVM: x86/mmu: Support shadowing NPT when 5-level paging is enabled in host) Signed-off-by:
-
Vitaly Kuznetsov authored
With the elevated 'KVM_CAP_MAX_VCPUS' value kvm_create_max_vcpus test may hit RLIMIT_NOFILE limits: # ./kvm_create_max_vcpus KVM_CAP_MAX_VCPU_ID: 4096 KVM_CAP_MAX_VCPUS: 1024 Testing creating 1024 vCPUs, with IDs 0...1023. /dev/kvm not available (errno: 24), skipping test Adjust RLIMIT_NOFILE limits to make sure KVM_CAP_MAX_VCPUS fds can be opened. Note, raising hard limit ('rlim_max') requires CAP_SYS_RESOURCE capability which is generally not needed to run kvm selftests (but without raising the limit the test is doomed to fail anyway). Signed-off-by:Vitaly Kuznetsov <vkuznets@redhat.com> Message-Id: <20211123135953.667434-1-vkuznets@redhat.com> [Skip the test if the hard limit can be raised. - Paolo] Reviewed-by:
-
Vitaly Kuznetsov authored
Commit 63f5a190 ("KVM: x86: Alert userspace that KVM_SET_CPUID{,2} after KVM_RUN is broken") officially deprecated KVM_SET_CPUID{,2} ioctls after first successful KVM_RUN and promissed to make this sequence forbiden in 5.16. It's time to fulfil the promise. Signed-off-by:
-
Vitaly Kuznetsov authored
hyperv_features's sole purpose is to test access to various Hyper-V MSRs and hypercalls with different CPUID data. As KVM_SET_CPUID2 after KVM_RUN is deprecated and soon-to-be forbidden, avoid it by re-creating test VM for each sub-test. Signed-off-by:
-
Sean Christopherson authored
Fully emulate a guest TLB flush on nested VM-Enter which changes vpid12, i.e. L2's VPID, instead of simply doing INVVPID to flush real hardware's TLB entries for vpid02. From L1's perspective, changing L2's VPID is effectively a TLB flush unless "hardware" has previously cached entries for the new vpid12. Because KVM tracks only a single vpid12, KVM doesn't know if the new vpid12 has been used in the past and so must treat it as a brand new, never been used VPID, i.e. must assume that the new vpid12 represents a TLB flush from L1's perspective. For example, if L1 and L2 share a CR3, the first VM-Enter to L2 (with a VPID) is effectively a TLB flush as hardware/KVM has never seen vpid12 and thus can't have cached entries in the TLB for vpid12. Reported-by:
Lai Jiangshan <jiangshanlai+lkml@gmail.com> Fixes: 5c614b35 ("KVM: nVMX: nested VPID emulation") Cc: stable@vger.kernel.org Signed-off-by:
-
Sean Christopherson authored
Like KVM_REQ_TLB_FLUSH_CURRENT, the GUEST variant needs to be serviced at nested transitions, as KVM doesn't track requests for L1 vs L2. E.g. if there's a pending flush when a nested VM-Exit occurs, then the flush was requested in the context of L2 and needs to be handled before switching to L1, otherwise the flush for L2 would effectiely be lost. Opportunistically add a helper to handle CURRENT and GUEST as a pair, the logic for when they need to be serviced is identical as both requests are tied to L1 vs. L2, the only difference is the scope of the flush. Reported-by:
-
Sean Christopherson authored
Flush the current VPID when handling KVM_REQ_TLB_FLUSH_GUEST instead of always flushing vpid01. Any TLB flush that is triggered when L2 is active is scoped to L2's VPID (if it has one), e.g. if L2 toggles CR4.PGE and L1 doesn't intercept PGE writes, then KVM's emulation of the TLB flush needs to be applied to L2's VPID. Reported-by:
-
